A warranty bond is the security that covers the phase after the job is finished: your bank stands in for defect claims your customer raises after acceptance of a building or other works. Anyone who signs a construction or plant-engineering contract knows the clause it comes from — after acceptance, some of the risk stays with the customer, for defects that only show up later. Instead of withholding part of your contract price, most customers ask for a warranty bond to cover it. For you as the contractor, that means full payment up front, but a deed that keeps tying up your guarantee facility for years afterward.
This guide explains what a warranty bond is, how it differs from a performance bond and a cash retention, how long it typically runs, what it costs you while it is outstanding, what “waiving the defence of prior enforcement” means in the bond wording, what happens if the customer actually calls it — and when and how you get the deed back once the defect period has run.
What is a warranty bond?
A warranty bond is a guarantee your bank issues on your company’s behalf to your customer, covering claims for defects that surface after acceptance of a building or other works contract. Legally, it is a suretyship under Section 765 BGB (the German Civil Code): the surety — your bank — undertakes toward the creditor of a third party (your customer) to answer for the fulfillment of that third party’s obligation. The statute explicitly allows a suretyship for a future or contingent obligation, which is exactly the case here: at the time the bond is issued, nobody yet knows whether a defect will even occur.
Economically, the warranty bond replaces a cash retention: instead of the customer withholding part of the contract price until the warranty period ends, they pay the full amount and take your bank’s bond deed instead. For you, that means full liquidity from the contract — paid for with guarantee commission instead of tied-up contract proceeds.
How it arises: drawn from your guarantee facility
A warranty bond is not a separate product — it is a single guarantee drawn against the guarantee facility your bank has already granted you. The process follows the same pattern as any other bond: you ask your bank to issue it, naming the beneficiary (your customer), the amount and the term the contract requires, and submit the bond wording the customer asks for. The bank pays out nothing at this stage — it only lends its creditworthiness. Only if the customer actually calls the bond does the bank become liable to pay the customer, and then claims the amount back from you.
Every warranty bond you issue ties up part of your facility until it is returned. Running several projects with different warranty periods at once makes a consolidated view worthwhile: total exposure = the sum of all outstanding guarantees, and free facility = facility granted minus total exposure. The bank mirror generator shows how much of your lines are already committed, on a single page.
How it differs: performance bond, warranty bond, cash retention
Project business uses several security instruments that are easy to confuse because they often appear in the same contract:
| Instrument | Covers | Runs until |
|---|---|---|
| Performance bond | Proper completion of the contract | acceptance |
| Warranty bond | Defects after acceptance | contractually agreed period, usually anchored to Section 634a BGB |
| Cash retention | Defects after acceptance (alternative to a bond) | paid out once the period expires |
The difference from a performance bond is timing: one covers delivery up to acceptance, the other covers defect-freedom afterward — which is why the warranty bond runs so much longer. The difference from a cash retention is the form of security: with a retention, the customer actually ties up your cash; with a bond, your bank only lends its creditworthiness — your contract proceeds are paid out in full.
LiquidityLens tracks your warranty bonds alongside performance and advance-payment guarantees — amount, deadline and total exposure in the same financial status.
See the Liquidity Control SprintHow long does a warranty bond run?
There is no statutory duration for a warranty bond itself. How long the deed stays valid is set by the bond wording and the construction contract — typically anchored to the statutory limitation period for defect claims under Section 634a(1) BGB (the German Civil Code): two years for works whose outcome consists of producing, maintaining or altering an object, and five years for a building.
The distinction matters: Section 634a BGB governs the limitation of defect claims against you as the contractor — not the bond’s duration itself. “Five years for a building” does not automatically mean “the warranty bond runs for five years.” In practice, the contracting parties anchor the security period to that limitation period, because defect claims generally cannot be enforced after it has run — but they are free to agree a different duration.
In German construction contracts, security arrangements are often additionally shaped by VOB/B — the standard terms and conditions for German construction contracts (Vergabe- und Vertragsordnung für Bauleistungen, Teil B), which independently governs the return and step-down of securities. VOB/B is not a statute; it only applies if the contract expressly incorporates it. Which period and which return mechanics apply in a given case is therefore always a matter of the specific VOB/B version agreed and the bond wording itself, not of any statute.
Legal note: This article explains the commercial context and the statutory basics in outline. It is not legal advice and does not replace it. Whether a specific security clause or a long bond term is enforceable in a given case — including under German standard-terms law (AGB-Recht) — is a legal assessment; consult counsel specializing in German construction and contract law for that.
What does a warranty bond cost?
There is no premium to pay when the deed is issued — the bank pays out nothing, so there is nothing to finance. What you pay is guarantee commission: a percentage per year on the amount guaranteed, charged for as long as the bond is outstanding. The rate itself is not a published tariff; it is negotiated with each bank and depends on your credit standing and the type of guarantee, so the only reliable figure is the one in your own facility agreement.
Two things follow from that for a warranty bond specifically. First, the term drives the cost more than the rate does: a bond anchored to a five-year limitation period carries five years of commission, where a performance bond ends at acceptance. Second, the cost does not stop when the defect period does — it stops when the deed comes back or an expiry clause bites. Everything between those two dates is commission on a risk the customer can no longer raise.
Set against that is what a warranty bond saves you. With a cash retention, part of your contract price simply stays with the customer until the period ends — money you have earned and cannot use. The bond buys that cash back at the price of commission. Whether that trade is worth it is a calculation you can only run with your own numbers: the bonded amount, the agreed term, your commission rate — and what the released cash is worth in your business.
Waived defences: what “selbstschuldnerisch” means
Warranty bond wordings almost always include the German term “selbstschuldnerisch.” The statutory default is different: under Section 771 BGB (the German Civil Code), a surety may refuse to satisfy the creditor as long as the creditor has not first attempted enforcement against the principal debtor — meaning you — without success (the defence of prior enforcement, Einrede der Vorausklage). Without waiving that defence, your customer would have to pursue you first before turning to the bank.
“Selbstschuldnerisch” means that defence is waived by contract. The bank then pays your customer without first having to enforce against you — that is a matter of the bond wording itself, not of Section 771 BGB. Combined with a bond payable “on first demand,” this means your customer gets paid before any dispute over the underlying defect is resolved. For you as the contractor, that is the point to scrutinize in the bond wording before signing — not whether the document is labeled “guarantee” or “suretyship.”
If the customer calls the warranty bond
A warranty bond moves no money until it is called. When the customer believes a defect has surfaced and you have not remedied it, they do not sue you first — they turn to the bank with the deed. What happens next is written in the wording, not in the statute: if the defence of prior enforcement is waived and the bond is payable on first demand, the bank pays the agreed amount without waiting for the question of who is right about the defect to be settled.
The money then comes back from you. The bank has lent its creditworthiness, not its cash, so once it pays out it claims reimbursement from you under your facility agreement. From one day to the next, a position that never appeared in your cash flow forecast becomes a payment out, at the full bond amount. Any argument about whether the defect was really yours happens afterwards, between you and the customer.
That is the practical reason to treat open warranty bonds as a liquidity item rather than a legal one. Two questions are worth answering before a call ever comes: which bonds are drawable on first demand, and what the largest single bond among them would do to your cash position in the week you have to reimburse it. The first is in the deeds; the second only shows if you keep the amounts in one place.
Getting it back: when and how
Whether a warranty bond ends automatically depends on its wording. If the deed carries an effective expiry clause, the bank’s liability lapses on its own once the agreed period passes — the only remaining step is confirming the bank has actually recorded the release against your facility. Without such a clause, or if the bond is open-ended, the facility is only freed once either the original deed is returned to the bank or the customer waives its rights under the bond in writing.
In practice that means: once the agreed period expires, write to the customer requesting the deed back, record the return in your own guarantee register, and verify the release on the bank’s next guarantee statement. This rarely happens by itself — without an active follow-up, an expired warranty bond often sits far longer than it needs to.
If the bond is never returned
If the return never happens, nothing changes about the exposure on its own: your guarantee facility stays tied up at the full bond amount, regardless of whether the defect period expired long ago. And the bank will usually keep charging guarantee commission on the amount — whether or not the customer will ever actually call it.
That is exactly where a forgotten warranty bond costs real money without showing up anywhere on a bank statement: the commission keeps running, and the blocked facility is missing when the next contract needs a performance or advance-payment bond. The more projects run in parallel with different warranty periods, the more likely exactly this gets overlooked.
For your planning: a register, not a footnote
A warranty bond moves no cash as long as it is not called — which is exactly why it easily falls out of view. Three things belong next to your cash flow forecast, not inside it: the expiry date of every individual bond, total exposure of all open guarantees against your facility, and the drawdown risk that becomes cash-relevant immediately if a bond is ever called.
That is what a guarantee overview delivers: one row per deed with beneficiary, amount, term and issuing bank, giving you total exposure and free facility at a glance. LiquidityLens keeps this register in a dedicated guarantees view inside the same financial status as your bank lines, receivables and 13-week cash flow forecast — so an expired warranty bond shows up from day one instead of surfacing on the next guarantee statement.
Weekly grid, cash in/out, low-point alert — ready to use.
FAQ
The most common questions about warranty bonds, answered below.