Duty to Defend Compared With Duty to Indemnify

An insurer's duty to defend is broader than its duty to indemnify. The duty to defend is normally triggered if any allegation in the complaint could potentially fall within coverage — even if the rest of the suit is plainly excluded, and even if the allegations turn out to be groundless. The duty to indemnify, meaning the duty to actually pay a settlement or judgment, depends on what the facts finally establish. An insurer can therefore owe a complete defense and, at the end of the case, owe nothing at all. Insurance is state law, and every rule described below varies from state to state.

1.Two promises in one policy

A standard liability policy makes two separate promises. One is to pay sums the insured becomes legally obligated to pay as damages because of covered injury. The other is to defend any suit seeking those damages. They are independent. Courts across the country describe the defense promise as a "litigation insurance" benefit that the policyholder bought in its own right, and they measure it by a much looser standard than the payment promise.

Duty to defend

Measured at the outset by what is alleged. If any allegation is potentially covered, the insurer must hire and pay counsel and fund the defense of the whole suit.

Duty to indemnify

Measured at the end by what is proved or agreed. It exists only for damages that actually fall inside the grant of coverage and outside every exclusion.

Because the duties are measured at different moments and against different material, the defense duty attaches earlier, lasts longer, and is harder for an insurer to escape. A California decision from 1966, Gray v. Zurich Insurance Co., is often cited for the proposition that the insurer must defend even suits that are ultimately without merit, because groundless suits still cost money to beat. That reasoning has spread widely, though each state has developed its own vocabulary for it.

2.The eight-corners comparison

The dominant method for deciding whether a defense is owed is to lay the complaint beside the policy and compare them. The four corners of the pleading are read against the four corners of the policy, which is why the test is usually called the eight-corners rule. Nothing else is consulted. The insurer may not rely on its own investigation to prove the claim is really uncovered, and the insured may not rely on facts it has not pleaded.

• The allegations are assumed true, however implausible they look.
• Doubts about the meaning of the pleading are resolved in favor of a defense.
• Ambiguity in the policy is generally construed against the insurer that drafted it.
• A label on a cause of action does not control; courts look at the conduct described.

A minority of states permit extrinsic evidence in defined circumstances. California, for instance, has long held that facts known to the insurer outside the pleading can create a defense obligation, a position associated with the state supreme court's 1993 decision in Montrose Chemical Corp. v. Superior Court. Texas, historically among the strictest eight-corners jurisdictions, recognized a narrow exception in 2022 in Monroe Guaranty Insurance Co. v. BITCO General Insurance Corp., allowing limited extrinsic evidence where the pleading is silent on a fact that decides coverage and the evidence does not touch the merits of the underlying suit. Neither approach is the national rule; both are examples of how far the same doctrine can drift between borders.

3.Suits that mix covered and uncovered claims

Most real lawsuits plead several theories at once — negligence alongside intentional conduct, contract alongside tort, statutory claims alongside common-law ones. The general rule is that a single potentially covered claim pulls the entire suit into the defense obligation. The insurer cannot defend the negligence count and leave the insured to fund the fraud count.

Payment is different. When judgment finally enters, the insurer indemnifies only the covered portion. That creates an allocation problem, and states resolve it differently: some place the burden on the insurer to prove which damages are uncovered, some place it on the insured to prove which are covered, and some ask the trial court to use special verdict forms so the split is visible on the record. Where the allocation cannot be untangled, several states resolve the ambiguity against the insurer.

4.Defending under protest, or refusing

An insurer that sees a coverage problem but cannot rule out a defense obligation has three practical routes. It can defend without qualification, which risks waiving coverage defenses. It can defend while reserving its rights, which is by far the most common course and is covered in detail in our entry on reservation of rights letters and what they preserve. Or it can refuse and file a declaratory judgment action asking a court to say no defense is owed.

Caution: Refusing a defense is the highest-risk option an insurer has. In many states a wrongful refusal forfeits control of the defense, obliges the insurer to reimburse reasonable defense costs, and may bind it to the outcome of a settlement the insured then makes on its own. In some states it also supports a claim for bad faith and extracontractual exposure beyond policy limits.

Declaratory judgment actions on coverage are frequently filed in or removed to federal court on diversity grounds, and the federal judiciary publishes general procedural information at uscourts.gov. A federal judge sitting in diversity still applies the coverage law of the relevant state, so the choice of forum changes procedure rather than substance. Background definitions of core insurance concepts are collected by Cornell's Legal Information Institute at its insurance law overview and its shorter insurance entry.

5.Can the insurer get defense costs back?

Suppose the insurer defends under a reservation of rights, wins a declaration that no coverage ever existed, and then asks the policyholder to repay what the defense cost. States are genuinely split. One group allows reimbursement on an implied-contract or unjust-enrichment theory where the reservation letter expressly claimed the right. Another group refuses, reasoning that the insurer bought the risk of an over-inclusive defense promise when it wrote a policy with a broad defense clause and cannot rewrite it after the fact. As of mid-2026 the split remains live, and the answer in a given case is usually determined by which state's law governs the policy.

Because the answer varies so much, the governing-law question can be worth as much as the coverage question itself. Policyholders with operations in several states often find that the same tender produces different outcomes depending on where suit is filed, an issue that overlaps with the drafting choices discussed in choice of law and forum selection.

6.Questions this raises

If the complaint pleads only intentional conduct, is a defense still owed?

Often yes. Many states look past the label to the conduct described, and a pleading of intentional harm frequently contains factual allegations that would also support a negligence theory. Some states go further and require a defense until the pleading forecloses every covered reading. Others enforce the pleading as written. Because the analysis is state-specific, the same complaint can produce opposite answers in neighboring states.

Does the duty to defend cover the cost of a counterclaim the insured wants to bring?

Usually not. The defense promise covers defending claims against the insured, not prosecuting the insured's own affirmative claims. Some policies and some state decisions treat a compulsory counterclaim that is inseparable from the defense as covered, and a few insurers agree to fund one when it will reduce overall exposure. Absent that, the insured normally pays for offensive work itself.

What if the plaintiff amends the complaint and removes the only covered claim?

The defense obligation is generally reassessed on the operative pleading. If an amendment removes every potentially covered allegation, the insurer may in many states withdraw prospectively, usually after written notice and often after giving the insured time to arrange substitute counsel. Withdrawing without notice, or in the middle of trial, exposes the insurer to argument that the timing itself caused harm.

Does a demand letter trigger the duty to defend before any suit is filed?

Under most standard forms the promise runs to a "suit," which typically means civil proceedings, and a bare demand letter is not enough. Many forms extend the definition to arbitrations and to alternative dispute proceedings the insurer consents to. Some environmental and regulatory contexts have produced decisions treating agency proceedings as suits. Read the definition in the specific policy rather than assuming.

7.Working through a tender

1. Tender at once, in writing. Send the complaint and every related demand to each insurer that might respond, not just the obvious one. Late tender creates its own problems, addressed in late notice and forfeiture of coverage.
2. Identify every policy year. Long-tail injury may implicate several years of coverage; the choice of trigger is examined in occurrence and claims-made policies.
3. Read the reservation letter closely. Note each defense reserved, each request for reimbursement, and any statement about who controls counsel.
4. Watch the pleading. Amendments change the analysis in both directions. Keep the insurer informed of each one.
5. Preserve the record. Keep defense invoices, correspondence, and settlement communications separate and organized in case the coverage question is litigated later.
6. Check your state's regulator. Complaint and market conduct resources are indexed through the National Association of Insurance Commissioners, and each state department can be located through USA.gov.

The practical lesson is simple. Treat the defense question and the payment question as two different fights, on two different timetables, decided by two different sets of facts — and never assume that a promise to defend implies a promise to pay.