Performance Bond: a Worked Example

A client awarding a building contract wants money it can reach if the contractor stops or fails. It could hold back cash. A bond gives it a promise from a stronger party instead, leaving the contractor's cash in the business. The bank charges a fee on the face amount for as long as the bond is out, and in exchange it has agreed to pay a sum it may then be unable to recover from the customer.

So the bank approves a bond much as it approves a loan, within a guarantee line and against the same security. The calling of a bond tends to coincide with the contractor's other troubles, and the bank that issued it is usually also the bank that lent it the overdraft. Capital rules count the promise as well: the Basel standardised approach applies a 50% credit conversion factor to transaction-related items such as performance bonds, and each regulator sets its own version of the table.

The wording decides what the beneficiary must show. An on-demand bond pays against a compliant demand, even if the contractor has not in fact defaulted. A conditional bond pays only once default under the contract, and commonly the loss it caused, is established. The label does not settle which is which: courts read the document as a whole, and English judges have held a guarantee described as payable on demand to be conditional. Market habits differ. In UK domestic construction, conditional bonds, often from insurers, are common and on-demand bonds less so. In Australia, unconditional undertakings from a bank or insurer are familiar security on commercial building contracts. In New Zealand both forms appear, and in 2024 the High Court restrained a call on a bond described as unconditional until an adjudicator had ruled, looking to the terms of the contract and the bond ([2024] NZHC 561).

Expiry is the other term a lender reads. A bond may end on a date, at an event such as practical completion, or only when the beneficiary releases it or returns the original. Many step down to a smaller amount at completion for the defects period.

A commercial builder, constructed for this page like every figure in it, wins a US$9.0m contract for a council library. The council requires a 10% on-demand bond, or US$900,000. The builder has a US$2.0m guarantee line at a fee of 1.5% a year, and an overdraft with US$600,000 drawn. The bank holds security it expects would realise US$1.0m. The percentages and amounts belong to the example.

The builder then tenders for a job needing a US$500,000 bond and finds it has US$150,000 of room. Nothing in its current work explains the shortage. The clinic finished eighteen months ago, and its bond has no end date, so it runs until the owner gives it back. Retrieving it would restore headroom to exactly US$500,000 and save a fee of US$5,250 a year (350,000 × 1.5%) on a job that is over, but the building has changed hands and the new owner cannot find the original.

A year later the builder runs out of cash with the library 70% built. The council serves a demand for US$900,000. The bank checks the demand against the bond, pays within days, and debits its customer.

The US$350,000 bond has come back to life as a risk, because a principal facing a collapsing contractor has every reason to look for what it can still call. Whether the new owner holds the benefit of the bond, and has a valid ground for a demand, depends on the wording, but the bank cannot assume it will not be asked.

Each instrument covers a different stage of a contract. A bid bond protects the party running a tender against a winner who will not sign. An advance payment guarantee returns money paid up front if the work is not done. A performance bond covers the delivery of the work itself, and a retention bond stands in for cash the client would otherwise hold back until defects are cleared.

An insurer's surety bond can do the same job as a bank's. For the contractor the difference is capacity: a surety line can leave bank facilities free, though a surety will often want an indemnity from the contractor or its owners. A standby letter of credit is the form more often seen when the beneficiary is overseas.

DEFINED IN ONE PARAGRAPH → · PRODUCT FILE: BANK GUARANTEE → · BANK GUARANTEE VS LETTER OF CREDIT →

Three readings recur. The first is capacity: a contractor's order book cannot grow faster than its bonding line, so a request for more bonds is often a growth question in disguise. The second is correlation. A lender adds bonds to drawn debt and asks what happens if the worst job goes wrong while others are still open, because a contractor's failure can bring several demands at once. The third is housekeeping. A register that ties each bond to a contract, an expiry and a named person at the beneficiary is how forgotten exposure is found. The bank may ask for cash cover where the customer's credit has weakened.

A property lender reads the same document from the other side. Funding a building, it may want the contractor's bond in place and in a form it can rely on, because the bond is part of what finishes the project if the builder does not.

DRAWING A DEVELOPMENT LOAN → · ON-DEMAND GUARANTEES →