Reading a Carrier Rate Filing: How to Spot a Tiered LCM

A PEO or MGA that quotes workers' comp without knowing each carrier's filed manual rate is guessing. An experienced underwriter has a feel for where a given carrier lands on a given class, but a feel is not a number, and the number is public. Every carrier writing voluntary-market workers' comp in a loss-cost state files its rate structure with the state insurance department, and most of those filings are searchable through SERFF Filing Access.

The problem is that a rate filing is not written to be read. It is a packet of PDFs assembled for a regulator, and the loss cost multiplier may be on page two of a memo, in an attachment's filename, or buried in a miscellaneous-values table. This post covers what is in a filing, where the LCM hides, how tiers work, and how to turn the filing into a rate you can defend.

What an LCM is, and what it is not

In an NCCI state, the rating bureau publishes advisory loss costs by class code: the pure cost of expected losses per $100 of payroll, with no provision for the carrier's expenses or profit. The carrier files a loss cost multiplier that converts the loss cost into a manual rate. Manual rate equals advisory loss cost times LCM. Everything else in the premium calculation, from the experience mod to schedule credits to premium discount, is applied to that manual rate.

The LCM is the carrier's expense and profit load. The Nevada Division of Insurance's LCM filing instructions describe the components a carrier must account for: acquisition costs including commissions, field supervision, collection, general expenses, taxes and assessments, and a provision for profit and contingencies that may be offset by anticipated investment income. The carrier shows an expected loss ratio, which is one minus the total expense and profit percentage, and the formula LCM that follows from it. Missouri's filing instructions state the arithmetic the same way: an LCM of 1.25 corresponds to a 20 percent expense load, because 1 divided by 0.80 is 1.25.

Two things the LCM is not. It is not the carrier's loss expectation; that is in the loss cost, and a carrier that wants to deviate from the bureau's loss costs must file and justify that separately. And it is not necessarily a single number.

Why carriers file tiers

A carrier writing a broad book has accounts it wants and accounts it will take. The preferred account has a safety program, a clean loss history, and a mod under 1.00; the non-standard account has a mod over 1.20 and a history that makes the underwriter nervous. Charging both the same load overcharges the first and undercharges the second, so carriers file multiple LCMs by tier and assign each account to a tier at underwriting.

Nevada's instructions make the mechanics explicit. Tiered rating is permitted to reflect expected differences in expenses and losses; for each tier the carrier must identify the business segments it contains, support any loss cost modification, explain any expense differential, and provide a separate expense worksheet. The instructions also say that a carrier that does not apply its tiers as filed and approved commits unfair discrimination. Tiers are not a marketing gimmick; once filed, the carrier is bound by them.

In our work building CarrierIQ, we have seen the same carrier file a single LCM in one state and three tiers in another. Tiers show up as named tiers (preferred, standard, non-standard), as separate companies under one group, as class-code-specific multipliers where one hazardous class carries a higher LCM than the rest, and as pre-computed rate pages with no LCM stated at all, where the multiplier has to be backed out by dividing a filed rate by the bureau loss cost for the same class.

For a PEO, the practical question is which tier its business lands in. A PEO master policy is usually priced at the standard tier, so a PEO comparing carriers should compare standard-tier manual rates unless a carrier has said otherwise.

What is actually in the filing

Open a workers' comp rate filing on SERFF and you will see a summary page with the tracking number, company name and NAIC code, filing type, submission and disposition dates, disposition status, and, on the detail view, the effective date. Below that is the list of attachments.

The attachments are where the content lives. A typical rate filing includes a cover letter describing what is changing, an explanatory memorandum with the rate-level analysis, an expense exhibit supporting the LCM, a miscellaneous values page listing the LCM, expense constant, minimum premiums, and any schedule rating maximum, and, depending on the state, rate pages with the resulting manual rate by class. Some filings add premium discount tables and waiver-of-subrogation charges.

The filing type matters more than people expect. LCM changes also appear in Rule filings and in combined Form/Rate/Rule filings, so a search restricted to the Rate type will miss them. Company-name searches have their own trap: SERFF paginates results, and a carrier with many filings in a state may have its newest one on a page nobody opened. Sorting by disposition date and checking every filing type is slower and correct.

Finding the LCM in the attachments

The LCM is rarely in a consistent place. We have found it in a single sentence in a memo, in a table on the miscellaneous values page with a tier label in front of it, in an attachment's filename, and omitted entirely in favor of a rate page.

A few reading rules help. Start with the miscellaneous values page if there is one; it is the regulator's standard summary and usually lists every tier. If the filing names more than one company, each may carry its own LCM. If a class-specific multiplier appears, note which classes it covers and treat everything else as the base. If the attachment is a scanned image, expect to read it by hand or through OCR, and double-check any digits that look wrong.

And read the effective date on the filing's detail page, not the submission date on the results list. A filing submitted in November may be effective the following March, and a rate computed from it before then is wrong.

Turning a filing into a manual rate

Once you have the LCM and the tier, the rate computation is one multiplication per class, but which loss cost you multiply is a date question.

For pricing a specific policy, the correct inputs are the most recent bureau loss costs and the most recent carrier LCM that were each effective on or before the policy's effective date. A loss-cost filing effective after the policy date is advisory for planning and cannot be used to price that policy. A carrier LCM filed two years ago and never updated is still the carrier's current LCM in a file-and-use state; carriers are not required to refile annually, and the absence of a new filing means nothing has changed.

A rate built from a two-year-old LCM is not wrong because it is old; it is wrong only if the carrier has since filed a new one that you missed. The right practice is to carry the effective dates of both inputs alongside every computed rate, flag the ones that are aging, and re-check SERFF on a schedule. Bureau loss costs in most states update annually, so a loss cost more than a year old deserves a look. A carrier LCM can reasonably go longer.

Take a hypothetical to make it concrete. Suppose a bureau publishes an advisory loss cost of 0.16 for clerical code 8810 and 3.96 for plumbing code 5183, effective March 1. A carrier's standard-tier LCM is 1.45, filed the previous year and still in force. The manual rates for a policy effective July 1 are 0.232 and 5.742 per $100 of payroll. If the carrier's preferred tier is 1.09, the same classes price at 0.174 and 4.316. For a plumbing contractor with $2 million in field payroll, the difference between the two tiers on that one class is roughly $28,500 of manual premium before any mod or credits, which is why knowing the tier is not optional.

Why stale filings are dangerous

A PEO that prices a client off last year's loss costs in a state that just took a large increase has underpriced the account by that increase. An MGA that builds a rate sheet from a carrier's old single-tier LCM after the carrier moved to tiers will quote the wrong number to every producer. A broker comparing two carriers on loss costs from different effective dates is comparing nothing. The defense is the same in every case: every rate carries its sources and dates, and the sources get re-verified before something goes wrong.

CarrierIQ ingests SERFF filings, extracts the LCM by tier, pairs it with the bureau loss costs effective for the policy date, and computes the manual rate per class per tier with both source filings and both effective dates attached. Rates past the staleness threshold are flagged rather than hidden, and when no current filing can be found for a carrier in a state, the module says so and links to the state's SERFF search rather than silently quoting an old number. California is handled separately, because the WCIRB publishes pure premium rates rather than NCCI loss costs and the LCM model does not apply.

None of that replaces reading the filing when it matters. It makes sure the underwriter is reading the right one.