If your company self-insures workers compensation in any U.S. state, you post collateral with the state’s workers compensation bureau or the department of labor. That collateral secures your obligation to pay future benefits on already-incurred claims, and it stays posted for the full tail of the program. For a large self-insured employer, that collateral obligation can run into tens of millions of dollars, tied up in surety bonds, letters of credit, or cash deposits, for decades after an accident year has closed.
Most finance officers encounter collateral as a number handed down by the bureau. “You need to post $14 million in collateral for 2026.” The calculation is usually opaque, the methodology varies by state, and the relationship between the collateral number and the actuarial reserve is sometimes tight and sometimes strangely loose.
This article explains how collateral works, how your reserve analysis drives the requirement (in some states directly, in others indirectly), and how to negotiate with the bureau when the collateral feels excessive relative to your actual exposure.
What collateral actually is
Workers compensation in the U.S. is a state-level system of mandatory benefits. Each state requires employers to assure payment of those benefits. For insured employers, the insurance carrier’s capital provides that assurance. For self-insured employers, the assurance comes from posted collateral.
Collateral serves two purposes. First, it protects injured workers: if the employer becomes insolvent, the state can draw on the collateral to pay ongoing benefits. Second, it protects the state itself: most states also maintain a guaranty fund that would cover claims if a self-insured employer defaults, and collateral reduces the risk the guaranty fund has to take on.
The amount of collateral is set by the state based on the employer’s estimated unpaid benefits, sometimes with a load for uncertainty, a margin for catastrophic losses, or adjustments for the employer’s financial strength.
How states calculate collateral
Four basic approaches exist across the states, with many variations.
1. Reserve-based, tied directly to the actuarial report
Some states require the self-insured employer to submit an actuarial report annually, and the collateral is calculated as the total unpaid losses from the report, sometimes with a loading factor. California is the most prominent example. The California Self-Insurance Plan (CSIP) sets collateral based on the actuarial estimate of total unpaid claims, with factors that load for adequacy.
In this approach, the actuarial reserve is directly the driver of the collateral number. A higher reserve estimate produces a higher collateral requirement.
2. Formula-based, using reported losses
Other states use formulas based on historical paid and reported losses rather than an actuarial projection. Florida, for example, historically used a formula tied to incurred losses from recent accident years. The actuarial reserve may be submitted but does not directly drive the collateral.
In this approach, your reserve and your collateral can diverge materially, in either direction. You may have a collateral requirement based on reported losses that is smaller than your actuarial reserve, or larger.
3. Credit-based, with actuarial support
A third approach blends formula and actuarial judgment with a credit-based adjustment. Strong-credit employers may be allowed to post less collateral; weaker-credit employers may be required to post more. New York’s Workers Compensation Board uses this model, combining reserves, payment histories, and financial strength reviews.
4. Negotiated, with regulator discretion
Some states leave the collateral determination largely to regulator discretion, with the actuarial report as one input among several (financial statements, bond ratings, industry benchmarks). In these states, the relationship between reserve and collateral is more negotiable, and a well-prepared employer with strong financials can often post less than the reserve.
Why the collateral usually exceeds the reserve
When you see your collateral requirement, it is often higher than your actuarial central estimate. The difference reflects several loadings that states typically apply.
Margin for adverse development. States assume your reserve might develop adversely. Loading of 10 to 25 percent on top of the central estimate is common.
ULAE. Unallocated loss adjustment expense is the cost of running your claims operation through the tail. Many states require this to be included in the collateral calculation even if your actuarial reserve is loss-only.
Future medical inflation. For the long tail of medical benefits on permanent disability claims, states sometimes assume a higher inflation rate than the actuary projected. This loads the reserve.
Catastrophic margin. Some states include a specific loading for the possibility of catastrophic events not yet reflected in the reserve. California’s CSIP historically has applied such factors.
Excess layer considerations. For employers with specific excess insurance, some states credit the excess placement and reduce collateral accordingly. For employers without excess, the collateral reflects the unlimited exposure to catastrophic claims.
Credit loading. Employers with weaker credit or volatile financial metrics see additional loading.
The practical consequence: collateral is typically 105 to 150 percent of the actuarial central estimate, depending on state and employer.
How your reserve analysis affects collateral
Three dimensions of the actuarial work have direct collateral implications.
Central estimate magnitude
The most obvious. A lower central estimate produces a lower collateral number in reserve-driven states. Efforts to tighten the reserve (by addressing case adequacy issues, by explaining prior adverse development, by improving data quality) can produce collateral reductions.
The inverse is true: an actuary that produces a central estimate above the prior actuary without a clear reason is costing you collateral capital.
Range construction
States that use ranges in their collateral calculation (some loading 25 percent above central; some loading to the 75th percentile of the range) are affected by the width of the range. A tight, well-supported range can reduce the loading.
A range that is padded without analytical support invites states to load aggressively.
Payment pattern
Some states calculate collateral based on expected runoff over a specific horizon (for example, expected payments over the next 10 years). The actuarial payment pattern therefore affects collateral. A pattern that pays out slower (longer tail) produces more collateral in such states because more losses remain unpaid at each future evaluation date.
An actuary who misjudges the payment pattern (through conservative assumption) costs you collateral over the life of the program.
The collateral lifecycle
Collateral is posted initially, adjusted annually as the program matures, and released only as claims actually pay out. The lifecycle matters because collateral ties up capital for decades.
Year 1: you begin self-insuring. You post initial collateral based on prospective estimates.
Years 2 to 5: collateral grows as accident years accumulate and unpaid losses build up. Each new year adds to the total obligation; each year’s partial payment reduces it modestly. Collateral typically peaks 5 to 10 years into a program.
Years 5 to 15: collateral plateaus and begins to decline as payments on older accident years outpace additions from new ones.
Years 10 to 30+: collateral declines slowly. The long tail of workers comp (permanent disability benefits, lifetime medical on catastrophic claims) keeps collateral above zero for decades.
After program termination: if you stop self-insuring, the collateral obligation continues for all accident years while self-insured. You continue filing, reserving, and posting collateral until all claims close.
Practical implication: the self-insurance decision is not just a decision about current-year economics. It is a decision about committing collateral capital for twenty-plus years.
Forms of collateral
States typically accept three forms of collateral.
Surety bonds: an insurance company guarantees your obligation for a premium. Surety is capital-efficient (you pay a premium, typically 0.5 to 2.0 percent annually, rather than locking up capital) but surety capacity is limited for large programs, and surety companies underwrite the financial strength of the self-insured. A downgrade in your credit can push surety premiums up materially or cause loss of surety capacity.
Letters of credit: a bank commits to pay the state if you default. LOCs are more broadly available than surety for large programs but tie up bank credit lines and carry a fee (typically 50 to 150 basis points annually).
Cash deposits or securities: you post cash or acceptable securities directly with the state. Most capital-intensive but sometimes required when other forms are unavailable or inadequate.
Most self-insureds use a mix. The mix should be reviewed annually against cost, capital availability, and credit capacity.
How to negotiate with the bureau
When the collateral number comes in and feels high relative to your exposure, four levers exist.
1. Challenge the actuarial reserve
If the collateral is directly driven by the actuarial reserve and you believe the reserve is overstated, the conversation starts with the actuary. A second opinion from an independent reviewer can sometimes produce a lower central estimate or a tighter range, which the state will accept as the basis for revised collateral.
2. Challenge the loading
If the state has applied a loading you think is excessive (for example, a catastrophic margin that does not reflect your actual exposure, or an adverse development loading that contradicts your program’s favorable development history), make the case in writing. States will often reduce loadings in response to evidence.
3. Provide credit support
If the state uses a credit-based model, strengthening the credit case (recent financial statements, bond rating improvement, guaranty from a stronger parent) can reduce collateral.
4. Restructure the program
If the collateral is driven by specific features (no excess insurance, exposure in high-benefit states, long-tail severity in medical), structural changes (adding excess, moving operations, closing high-cost states) can reduce the collateral obligation over time. This is a long-horizon lever, not a quick fix.
What states care about in the actuarial report
If you are in a state that relies heavily on the actuarial report, the report itself is the most important input to the collateral conversation. States look for:
- A credentialed signing actuary with relevant experience.
- Clean data reconciliation to the state’s incurred loss database or your claim system.
- Methods appropriate for the program’s age and size.
- Documented diagnostics that explain any unusual patterns.
- A range with a clear central estimate, so the state can apply its loading rules unambiguously.
- Year-over-year consistency or, if methodology changes, clear documentation of why.
A report that meets these criteria gives the state fewer reasons to load beyond its baseline. A report that does not invites the state to take the high end of the range, add margins, and push you for additional collateral.
Budget implication
For a multi-state self-insured program, collateral is frequently the largest balance sheet commitment related to the program, often multiples of the annual claim spend. Budget for collateral as a capital commitment, not just a current-year expense. Model the collateral runoff over 20 years as part of any decision to self-insure, exit, or restructure.
Related reading
- What a Reserve Review Should Cost: fee context for reserve work that directly drives collateral.
- Workers Compensation IBNR for Self-Insured Employers: the mechanics of the reserve the state is relying on.
- When to Get a Second Opinion on Your Reserves: how an independent reserve view can support negotiating a revised collateral number.
If you are facing an increased collateral requirement and want an independent actuary to support a counter-proposal, the hire an actuary directory is in development. Join the waitlist there.