What a hold harmless clause does
A hold harmless clause answers one question: when a defined loss lands, who is legally shielded from bearing it. The clause creates a contractual shield that either prevents one party from pursuing the other, requires that party to cover the other’s losses, or both.
In practice, hold harmless language carries two overlapping functions:
- Exculpatory function: the promisor agrees not to hold the protected party liable for its own losses, giving up claims it might otherwise bring. This is the “release” side of the clause.
- Reimbursement function: the promisor agrees to cover losses, including third-party claims, that the protected party incurs. This overlaps heavily with indemnification, which is why the two phrases so often travel together as “indemnify and hold harmless.”
Because courts do not treat “hold harmless” as a fixed term of art, its reach depends entirely on the surrounding words. Drafters commonly describe three levels of coverage, borrowed from construction practice:
- Broad form: the promisor covers losses even when they are caused entirely by the protected party’s own negligence.
- Intermediate form: the promisor covers losses caused in part by the protected party, but not those caused solely by that party.
- Limited (comparative) form: the promisor covers only losses attributable to its own fault, in proportion to that fault.
The form you choose is the single most important decision in the clause, because it decides whether a party can be shielded from the consequences of its own conduct. A well-built provision also fixes the trigger (what event sets it off), the covered losses (which money is included), the procedure (notice, defense, and settlement), and the limits (caps, carve-outs, and survival). Leave any of these vague and the clause underperforms exactly when it is needed.
Drafting example
Hold harmless. The Contractor shall indemnify and hold harmless the Owner and its officers, employees, and agents (the “Protected Parties”) from and against any claims, damages, liabilities, and reasonable attorneys’ fees (collectively, “Losses”) to the extent arising out of the Contractor’s performance of the Work or breach of this Agreement. This obligation does not apply to Losses to the extent caused by a Protected Party’s own negligence or willful misconduct.
Read the load-bearing phrases. “To the extent arising out of” ties the shield to causation and supports proportional allocation rather than all-or-nothing exposure. The closing sentence is a carve-out that keeps the Contractor from paying for the Owner’s own fault, which makes this a limited (comparative) form clause. Pairing “indemnify and hold harmless” supplies both the reimbursement duty and the release, so the Owner is protected whether a loss is a third-party claim or a direct one. Every phrase here is a lever you can tighten or loosen in negotiation, and each deserves a deliberate choice.
What US law says
Hold harmless clauses are governed by state contract law, and the rules vary meaningfully from one jurisdiction to the next. Several themes recur.
First, courts read these clauses narrowly and against the drafter. To shield a party from the consequences of its own negligence, many states demand clear and specific language, and some apply an “express negligence” rule requiring that the intent be stated in unmistakable terms. Generic “any and all claims” wording may not be enough.
Second, many states restrict these promises by statute. Anti-indemnity statutes, common in construction, limit or void broad form clauses that make one party hold another harmless for that other party’s sole negligence. Because these statutes differ by state and by contract type, a clause that is enforceable in one state may be partly void in another.
Third, public policy sets outer bounds. Courts often refuse to enforce a hold harmless promise that purports to release a party from its own fraud, gross negligence, or willful misconduct, and consumer-facing releases face heightened scrutiny.
Fourth, when the clause functions as a release signed by a consumer (a gym waiver or a recreational activity form, for example), enforceability can turn on how conspicuous and clear it is and whether the activity implicates a strong public interest.
Because enforceability depends on the governing law you select and the state where a dispute lands, confirm the operative rules for your jurisdiction before relying on any specific wording. This is general legal information, not legal advice.
Common mistakes to avoid
- Treating “hold harmless” as self-defining. The phrase has no fixed meaning. Spell out the trigger, the covered losses, and whose negligence is included rather than trusting the label.
- Reaching for broad form by default. A clause that shields a party from its own sole negligence is the hardest to enforce and, in many states, void by statute. Match the form to what the law and the risk actually justify.
- No carve-out for the protected party’s own fault. Without a “to the extent” or comparative-fault carve-out, one side can end up paying for losses the other caused.
- Uncapped exposure. Hold harmless duties frequently sit outside the contract’s limitation of liability, so a single covered claim can exceed the deal value. Decide deliberately what is capped and what is carved out.
- Silent on defense and settlement. If the clause does not say who controls the defense and who must approve a settlement, you can lose leverage over a claim you are funding.
- Forgetting survival and insurance. If the promise does not survive termination, it may expire before latent claims surface, and a promise not backed by insurance may be worth little when a loss actually lands.
When it matters most
Hold harmless clauses earn their keep wherever one party’s activity can generate losses that reach the other: construction and contracting (jobsite injury and property damage), commercial leases (a landlord shielding itself from tenant-side claims), events and venue rentals, equipment rentals, recreational and fitness businesses, and service arrangements where a vendor works on a customer’s premises or systems. They also appear throughout supply, licensing, and logistics agreements. As a rule of thumb, the greater the chance that outside claims or on-site losses will arise, the more the exact wording is worth, and the more time it deserves at the table.
A hold harmless clause is only as strong as your ability to track and enforce it across a portfolio of agreements. Forms, carve-outs, caps, survival periods, and insurance requirements are easy to draft and easy to lose in a shared drive. A CLM platform like Pactolane keeps every executed contract in one repository, and PactAI can surface hold harmless terms in a multilingual executive summary, score broad or uncapped language against your compliance playbooks, and flag where one contract’s promise conflicts with another. The clause protects you on paper; disciplined contract management is what turns that protection into something you can rely on when a loss actually lands.
Related clauses
Frequently asked questions
What is the difference between a hold harmless clause and indemnification?
A hold harmless clause is usually read as a promise not to hold the other party responsible for certain losses, while indemnification is a promise to reimburse those losses after they occur. The two overlap so heavily that contracts routinely combine them as "indemnify and hold harmless," and some courts treat them as largely equivalent. Because interpretation varies by state, define exactly what is covered rather than relying on the labels.
Can a hold harmless clause cover a party's own negligence?
A hold harmless clause can shield a party from its own negligence only if the language is clear enough to say so. Many states apply an "express negligence" rule that requires the intent to be stated in unmistakable terms, and generic "any and all claims" wording may not qualify. In some sectors, such as construction, anti-indemnity statutes limit or void these broad promises entirely.
What are broad, intermediate, and limited form hold harmless clauses?
The three forms describe how far the promisor's protection extends based on whose fault caused the loss. A broad form covers losses even when the protected party is solely at fault, an intermediate form covers losses the protected party caused in part, and a limited (comparative) form covers only the promisor's own share of fault. Broad form is the hardest to enforce and is restricted or void in many states, so choose the narrowest form that fits the risk.
Should a hold harmless clause be mutual or one-way?
Whether a hold harmless clause should be mutual or one-way depends on where the risk actually sits. A one-way clause makes sense when only one party's activity can generate the relevant losses, while a mutual clause is fairer when both sides create comparable exposure. Match the structure to the real allocation of risk rather than accepting whatever the first draft proposes.
Does a hold harmless clause survive termination of the contract?
A hold harmless clause survives termination only if the contract says so. Because covered claims can surface long after an agreement ends, the promise should be listed in a survival provision with a clearly stated survival period. Confirm that the period lines up with the relevant statutes of limitation for your jurisdiction.
In the same family
Not to be confused with
The comparison that sets this clause apart from a neighbouring concept.
On the same topic
Other pages closely related to this one.