Excess vs umbrella, properly distinguished

Both products sit above primary policies and pay claims that exhaust them. The difference is breadth: an umbrella can drop down to fill primary gaps, an excess policy almost never does.

An umbrella policy is broader than the primary policies it sits over. It provides excess limits when underlying coverage applies, and in some cases it acts as primary coverage for losses the underlying policies do not address (subject to a self-insured retention, usually $10K to $25K). An excess policy is narrower. It follows the form of underlying coverage, picks up only after underlying limits are exhausted, and grants nothing the primary did not already grant. The market uses the words loosely. A "lead umbrella" written above $1M GL and $1M auto is in practice the same animal as a lead excess. The distinction sharpens once you climb above the lead and into the excess tower proper.

Why both exist at all

Primary GL is sold with $1M / $2M limits as a near-universal market default. That is enough for routine claims and entirely insufficient for a serious bodily injury verdict, a multi-fatality auto loss, or a products case with class certification. Buyers need limits an order of magnitude larger. Carriers, however, are unwilling to write a single $25M or $100M policy on most accounts because the volatility is too great for one line to absorb. The solution is to slice the limit into layers and spread them across multiple carriers, each taking a sliver of the total exposure. The result is a tower.

Tower mechanics: attachment, layers, ventilation

The tower is a stack of policies, each defined by an attachment point (where it starts paying) and a limit (how high it pays before exhausting). Each layer is its own contract with its own carrier and its own price. Together they aggregate to the buyer's total limit purchased.

Ventilation

A "vented" layer has a gap below it. If the tower above runs $25M xs $5M but underlying is $1M, the buyer has a $4M gap from $1M to $5M that is uninsured (or self-insured). Ventilation is sometimes deliberate (the buyer wants to retain the layer as self-insured because it knows its loss profile), and sometimes forced (no carrier would price the $4M xs $1M layer at a number the buyer would accept). Excess underwriters check for ventilation on every account because an unventilated tower is genuinely safer than a vented one. A loss reaching the excess layer means the buyer also retained the gap, which is a credit signal.

Quota share within a layer

A single $15M layer is often split across multiple carriers. Three carriers might each take a one-third quota share of the layer, paying pro rata on any loss that pierces it. This is how the market spreads volatility within a layer the same way the tower spreads it across layers. The quota share carriers all sit on the same paper at the same attachment, just with smaller individual exposures.

Follow-form vs manuscript wording

Excess policies above the lead are almost always follow-form. The policy says, in effect, "we cover what the underlying policy covers, with the same exclusions, definitions, and conditions, except as modified herein." The "except as modified" clauses are where the underwriter actually does work.

Schedule of underlying

Every excess policy attaches its schedule of underlying as an endorsement or exhibit. The schedule lists the primary policies the excess sits over, including carrier, policy number, effective dates, and limits. If the schedule is wrong, the policy is wrong. Underwriters spend real time reconciling the schedule against the actual primary policies because brokers commonly transcribe limits incorrectly, miss endorsements, or list a placeholder primary that has not yet been bound.

Drop-down and erosion

Two mechanics let an excess policy respond differently than expected.

Drop-down

If primary insurance is exhausted by losses (not by cancellation, not by insolvency unless the wording specifically responds to insolvency), the excess drops down to act as primary. Drop-down is the buyer's protection against having a thinning tower over a long claim year. If three large GL losses chew through the $2M general aggregate by August, the excess drops down and acts as primary for any loss occurring after that. Most excess policies do this. Most umbrellas do this and also drop down for losses the primary did not cover at all (subject to the umbrella's SIR).

Erosion

Erosion is the gradual consumption of the aggregate by paid losses. Underwriters track erosion mid-term on aggressive accounts because it changes the shape of the tower. A $5M aggregate eroded by $3M of paid losses is now functionally a $2M aggregate, and the excess underwriter wants to know before renewal because the next loss could pierce them earlier than the original schedule suggested.

Self-insured retentions and SIR mechanics

An umbrella policy may sit over a self-insured retention rather than a fronted primary. The buyer self-insures the first $250K, $500K, $1M, or more, and the umbrella attaches above that retention. SIRs are common in middle market and large commercial accounts that have stable loss experience and want to retain expected losses to avoid paying the carrier's loss-cost margin on predictable claims.

Aggregate SIRs vs per-occurrence SIRs

A per-occurrence SIR applies separately to each loss. A $250K SIR on an account with three $200K losses means the buyer retains all three, $600K total. An aggregate SIR caps total annual retention. A $500K aggregate on the same three losses caps the buyer at $500K and pushes the third loss into the umbrella for $100K. Aggregate SIRs are more buyer-friendly and more common on large accounts.

Defense within the SIR

Defense is usually inside the SIR. Buyers manage their own defense (with their own counsel, on their own clock) up to the retention. This is why TPA selection matters so much for SIR-bearing buyers. The TPA defends within the SIR. Once the SIR is exhausted, the umbrella carrier takes over defense. The handoff is one of the messier moments in casualty claims management.

Nuclear verdicts & social inflation

A verdict of $10M or more is "nuclear." The frequency of nuclear verdicts has reset the economics of US casualty over the last decade. This is the macro story behind why excess limits got expensive and why carriers got selective.

The drivers are well documented. Plaintiff bar consolidation has created firms with the capital to fund multi-year complex cases. Third-party litigation funding (TPLF) lets outside investors finance cases in exchange for a piece of the recovery, removing the cash constraint that historically capped plaintiff strategy. Reptile theory and similar trial techniques have moved jury anchors. Anti-corporate sentiment, post-2008 and post-COVID, has shifted public attitudes toward defendants. Specific industries (trucking, healthcare, opioid manufacturing, residential construction) have been disproportionately hit.

Venue concentration

Verdicts cluster geographically. Florida, Georgia, California, New York, Texas, Pennsylvania (especially Philadelphia), Illinois (Cook County) and Louisiana account for the majority of nuclear verdicts. Underwriters look at exposure by venue, not just by class. A trucking fleet that runs 60% of its miles through I-95 in Florida is priced very differently than one running through New England.

Pricing logic: ILFs, layering, severity curves

Excess pricing relies on Increased Limit Factors (ILFs) and severity distributions. The primary $1M layer has a known burn rate. Each additional million sits further into the tail of the severity distribution and gets cheaper per dollar of limit, until the layers get high enough that they hit the social-inflation tail and prices reverse.

The standard mental model

• Lead $4M xs $1M. Most expensive per dollar of limit. Sees the most losses. Subject to most adverse development.
• $10M xs $5M. Cheaper per dollar. Many accounts never pierce $5M.
• $15M xs $25M. Cheaper still on standard accounts. But severity-driven on tough accounts: trucking, residential GL, hospitality with assault, K-12, certain healthcare.
• $50M xs $50M and above. Pricing depends almost entirely on whether the account has nuclear-verdict exposure. Bermuda and London absorb most of this.

Pricing inputs that move the number

1. Class of business and venue mix (the two biggest single inputs above $5M)
2. Five-year loss history with attention to large losses, severity outliers, and reserve development
3. Underlying limits adequacy (a $1M primary feels different to an excess underwriter than a $2M primary)
4. Risk management: fleet safety, contractor controls, premises security, employment practices
5. Schedule of underlying integrity
6. Tower shape (vented vs unvented, primary carrier solvency, peer carriers in the tower)

How underwriters read a submission

An excess underwriter receives a submission and works through it in a roughly fixed order. The order matters because the early questions can decline the risk before the underwriter spends time on the rest.

The reading order

1. Class and venue. Is this even within appetite? A long-haul trucking submission with 80% Florida and Georgia miles is declined in five minutes by most carriers.
2. Loss history. Five years of loss runs by line. Are there severity outliers? Reserve development? A pattern?
3. Underlying program. What is the buyer attaching to? Is the schedule clean? Are limits adequate to the exposure?
4. Operations and revenue. What does the buyer actually do? Where? Has anything changed materially since last renewal?
5. Submission completeness. Application, ACORD 125/126, supplementals, drivers list (auto), payroll by class (WC), revenue by state, additional questionnaires for tough classes.
6. Tower context. What does the rest of the tower look like? Who is leading? Where is the buyer trying to attach the next layer?
7. Pricing. Only after all of the above.

The ideal submission pack

Where IDP earns its keep

Excess casualty submissions are document-heavy. A typical submission package runs 80 to 250 pages across applications, supplementals, schedules of underlying, loss runs from multiple prior carriers, narratives, and exhibits. The underwriter's job before pricing is largely a comprehension job: get the structured fields out, normalize the loss runs, reconcile the schedule, flag the issues.

The five excess workflows where Indico routinely shows up

• Schedule of underlying reconciliation. Pull the SOU off the application, pull the actual underlying policy details, flag gaps and inconsistencies. This is the highest-value extraction job in excess.
• Loss run normalization. Five years across three or four prior carriers, all in different formats. Normalize to one schema with claim type, date of loss, paid, reserved, status, cause.
• Submission triage. Appetite check against class, venue, limits, loss history thresholds. Route within seconds of intake.
• Application completeness. Compare what the broker sent against what the carrier needs. Generate a structured request-for-information for the broker.
• Renewal analysis. Compare expiring vs renewal terms across the entire tower. Highlight changes in attachment, limits, exclusions, premium movement.