Power performance letters of credit

A large-load interconnection is expensive for the utility long before it is useful to anyone. Capacity is reserved, studies are run, long-lead equipment is ordered, and in many cases construction begins on upgrades that exist for one customer and have no other purpose. The utility is spending against a request.

What it is protecting against is therefore not non-payment for something delivered. It is a customer that stops performing, stops paying its share of upgrade costs, or ceases to exist after the utility has committed to work that cannot be redirected. The exposure is to abandonment, not to invoice risk, and it runs for years.

Hence a collateral requirement, and hence the instrument. In this relationship the developer or load customer is the applicant — the party that asks a bank to issue the credit and reimburses the bank if it is drawn. The utility or grid operator is the beneficiary. The instrument is almost always a standby credit: an undertaking by a bank to pay on presentation of specified documents, drawn only if the customer fails to perform.

This is a different relationship from the one described in [the LC-backed equipment facility](lc-backed-equipment-facility). There the beneficiary is a supplier or a financier, and the credit substitutes a bank's promise to pay for the buyer's own — payment risk on goods. Here the beneficiary is a regulated counterparty that will not extend service without security, and the credit is collateral for performance and cost obligations under a tariff or a service agreement. Same instrument, different beneficiary, different trigger, different job.

Almost everything below follows from the identity of the beneficiary. A utility is not a commercial counterparty negotiating a bilateral document. It generally publishes the form it requires, applies it uniformly, and treats the collateral rules as part of the tariff rather than as terms. The developer's real negotiation is not with the beneficiary about the instrument; it is with a bank about issuing it, and with itself about what that costs.

The amount is rarely a matter of judgement. It is generally derived from a formula or a schedule, and the practical work is anticipating it rather than arguing with it.

The usual drivers:

The size of the request. The requirement scales with the capacity being reserved, because that is what the utility is holding out of service for someone else.

The cost attributable to the customer. Where network upgrades are being built for a specific request, the security is generally sized against the utility's exposure to that work — the money it will have spent if the customer walks away.

The stage of the process. This is the most important mechanic and the one most often planned for too late. Security is not a single posting. It steps up as the position advances: a deposit at application, more at study or agreement, and the largest increase generally at the point the utility commits to construction. A developer that has budgeted for the first step has budgeted for the smallest one.

The customer's own credit standing. Many tariffs give rated entities an unsecured allowance — an amount of exposure the utility will carry against the customer's own standing without collateral. A project entity formed for one site has no rating and no allowance, so the whole requirement is collateralised.

The length of the commitment. Security is held for the period the utility is exposed, which typically runs from the point of commitment through energization and, under some service agreements, into the operating term.

Two consequences are worth stating plainly. The first is that this is a liquidity commitment made years before revenue, sized against someone else's construction programme, at the stage of a development when capital is scarcest. The second is that it compounds across a pipeline: a developer advancing positions in several markets holds several of these at once, none of which is released early, which is a portfolio problem rather than a project one.

Tariffs and service agreements generally permit more than one form of security, and the choice is a genuine structuring decision rather than an administrative one. It is usually made on liquidity rather than on headline cost.

The instrument's mechanics are where developers lose money, and they are mechanical enough to work through in advance.

The validity period is shorter than the commitment. Banks issue contingent exposure for defined windows and review it when each comes up. The utility's exposure runs to energization and beyond. The gap between those two periods is the structural weakness, and it is the same mismatch that governs any letter of credit — set out in general terms at [letter of credit](/learn/letter-of-credit).

Evergreen provisions move the risk rather than removing it. Utility forms almost always require automatic renewal for successive periods unless the issuer gives notice of non-extension, coupled with the beneficiary's right to draw the full amount on receiving that notice. Read from the beneficiary's side this is protection. Read from the applicant's side it is a trigger: a decision by its own bank to stop renewing causes an immediate full drawing, which converts a contingent obligation into funded debt at precisely the moment the bank has concluded it no longer wants the exposure.

Drawing conditions are written to be easy. A utility form typically requires a sight draft and a signed statement that the customer has failed to perform. There is generally no requirement to prove the underlying default, and the credit's independence means a dispute about whether the customer actually defaulted does not stop payment. The correct reading is not that this is unfair; it is the point of the instrument. The practical implication is that the money leaves first and the argument happens afterwards, against a bank the applicant now owes.

Step-downs have to be automatic or they will not happen. As upgrade work is completed and obligations are discharged, the amount required should reduce. Whether reductions occur on a stated event or require the beneficiary's agreement is one of the few genuinely variable points, and over a long development it is worth a great deal of carried liquidity.

Release is a process, not a date. The obligation may fall away at energization, at completion of the upgrades, at the end of a stated service period, or when the customer qualifies for an unsecured allowance. None of that reduces the bank's exposure until the beneficiary returns the original instrument or issues a release. Until it does, the line stays used and the cover stays posted. This is the least interesting failure on the page and one of the most common.

When the requirement outgrows one bank. A developer holding several positions holds several credits simultaneously, all outstanding, none released early. Past some point that exceeds what a single institution will carry as contingent exposure to one group. The available answers each have a cost: a facility provided by a group of banks under one agreement and one agreed form, which standardises the mechanics but takes time to put in place; separate bilateral lines per position, which is faster and multiplies the number of expiry dates somebody has to diarise; a confirmation from an institution the beneficiary accepts, where the natural issuer is not on its list; or cash-collateralised issuance, which reintroduces exactly the liquidity cost the instrument was chosen to avoid. The structural point is that at that scale, issuance capacity becomes a constraint on development speed — positions can only be advanced as fast as security can be arranged behind them, which is a capital-planning question rather than a treasury one.

The recurring ways this goes wrong:

• A non-extension notice triggers a full drawing. The bank declines to renew, the beneficiary draws, and the applicant owes the full amount immediately and generally on secured terms. The mitigation is to know each credit's renewal decision date well before the bank reaches it, and to have an alternative issuer identified rather than sourced under pressure.
• The form was read too late. The required wording is usually an attachment to a tariff or a service agreement, not a negotiable draft. A developer that assumes it can amend the drawing conditions discovers that the only negotiation available is with the issuing bank, which is being asked to accept a form it did not write.
• The applicant is the wrong entity. A project entity formed for one site is generally not a bank customer on its own, so the parent applies — and the layers the structure was separating are quietly re-coupled. That is the same problem, approached from the other end, as [who holds the interconnection position](who-holds-the-interconnection-position). Contingent obligations of this kind may also be restricted by covenants in the parent's other arrangements.
• Step-downs that require consent never happen. Amounts sized against a construction estimate remain outstanding long after the work they secured is complete, because nobody has an obligation to reduce them and nobody chased it.
• The release is never actioned. The project energizes, the obligation is discharged, and the credit stays outstanding because the original was never returned. The cost is invisible on the project and real on the line.
• Line capacity becomes the binding constraint. A pipeline advances faster than security can be arranged behind it, and a position is lost or deferred not because the site failed but because there was nothing left to post.
• A guarantee was accepted instead, and the rating moved. Where a parent guarantee satisfied the requirement, a downgrade or a failed net-worth test converts it into a demand for cash or a credit, generally on short notice and generally in conditions where neither is easy to arrange.
• The security is treated as a development cost rather than a capital structure item. It is neither. It is a multi-year contingent obligation of a named entity, and deciding late which entity carries it produces the same misallocation as deciding late who holds the position it secures.

Continuum structures and arranges transactions in which these instruments are posted and coordinates the parties to them. It does not issue, confirm, advise or hold letters of credit, is not a bank, a broker-dealer or a direct lender, and does not hold client funds.