The Regulator Rejected a D&O Rate Filing Because Excess Layers Used Bermuda Benchmarks

In mid-2025, a state insurance department rejected a D&O rate filing for a new market entrant because the top $10 million excess layer was priced using Bermuda market benchmarks instead of domestic loss experience. The decision was not a routine deficiency letter—it triggered a renegotiation of rates across all layers of the tower and warned the broader market: excess layers are not exempt from rate adequacy scrutiny. For actuaries and underwriters who have long treated Bermuda indices as a convenient shortcut, the ruling was a reminder that regulatory standards are jurisdiction-specific, and that a benchmark is only as good as the data behind it.

The rejection did not happen in a vacuum. The carrier had submitted a multi-layer D&O program with a $25 million total limit. The primary layer of $5 million was priced using the carrier's own U.S. claims data, which was credible due to a decade of experience writing small-to-medium public companies. The first excess layer of $10 million used a blend of the carrier's data and industry loss ratios from U.S. sources. The top excess layer of $10 million, however, was priced using a Bermuda market index that aggregated claims from multiple jurisdictions. The regulator flagged this inconsistency immediately.

Why a D&O Filing Was Rejected: The Bermuda Benchmark Problem

Bermuda excess layer pricing has historically diverged from the U.S. admitted market. Bermuda-based carriers often write high-excess layers—typically the top $5 million to $10 million of a $25 million tower—using global loss indices that blend claims from multiple jurisdictions. The filing in question used a Bermuda index rate for the top $10 million layer, applied a 40% loading for jurisdictional risk, and submitted the resulting premium as part of a multi-layer program.

The regulator rejected the filing on the grounds that the rate lacked support from U.S. loss experience. The index, while widely used in Bermuda, was not calibrated to the frequency or severity of U.S. securities lawsuits, which differ materially from global patterns. The regulator cited a lack of domestic claims data for the excess layer and required the carrier to refile with U.S.-only loss triangles.

The rejection had immediate effects. The carrier had to renegotiate rates across all layers of the tower, as the primary and lower excess layers had been priced relative to the top layer. The delay pushed back the carrier's market entry by several months and increased overall program cost by an estimated 10–15% for the top layer.

The decision set a precedent that other states are now watching. Several insurance departments have informally indicated they will scrutinize excess filings that rely on non-U.S. benchmarks, particularly for D&O and professional liability lines where litigation risk is jurisdiction-specific. The Bermuda benchmark problem is now a regulatory focus.

How Excess Layers Are Priced: Benchmarks, Not Experience

Primary D&O layers are typically priced using the carrier's own loss experience, adjusted for exposure and trend. Rate filings for these layers include detailed loss triangles, frequency and severity distributions, and credibility-weighted projections. The actuarial standard is established: rates must be adequate, not excessive, and not unfairly discriminatory, based on the insurer's own data or credible industry data.

Excess layers, especially those above $10 million, face a data problem. Claims that penetrate to these levels are rare—often fewer than one per thousand policies. Carriers lack credible own-experience data, so they turn to external benchmarks: reinsurance indices, modeled rates from catastrophe modelers, or market surveys like the Bermuda D&O index. These benchmarks aggregate claims from multiple carriers and jurisdictions, but they are not designed for U.S. regulatory rate filing.

The Bermuda market benchmark, for example, draws on claims from London, Bermuda, and a handful of other non-U.S. markets. U.S. securities litigation is unique—plaintiffs' bar activity, class-action certification rates, and settlement amounts differ significantly from other regions. A global blended rate may understate or overstate U.S. risk, depending on the mix. The 40% loading applied in the rejected filing was intended to adjust for this, but the regulator deemed it arbitrary because it was not derived from a transparent data analysis.

The gap between modeled risk and actual claims data is an ongoing challenge. Reinsurers and excess carriers rely on these benchmarks because they have no better alternative. But regulators require that rates be based on loss experience that is relevant to the policies being written. When a filing uses a benchmark that has no direct connection to the insured exposure, the rate adequacy argument becomes difficult to sustain.

To illustrate the data challenge, consider a hypothetical but realistic scenario: a carrier writing D&O for mid-cap technology companies. Its primary layer may have 50–100 claims over five years, enough for credible loss development. The top excess layer, however, may have zero claims in the same period. The carrier then turns to a Bermuda index that includes claims from financial institutions in London and Bermuda—companies with different risk profiles and different legal environments. The index may show a loss cost of $0.50 per $100 of premium, but the U.S. technology sector's actual loss cost could be $0.80 or $0.30. Without adjustment, the rate is a guess.

The Regulatory Rule That Caught the Filing

The regulatory rule that caught the filing is rooted in the standard for rate adequacy. Under the NAIC rate filing guidelines adopted by most states, a rate must be "adequate to cover expected losses and expenses." The key word is "expected"—the insurer must demonstrate that the rate is based on a reasonable estimate of future claims, using data that is appropriate for the risk being insured.

The state insurance department cited Section 4 of its rate law, which requires that rates be "based on the insurer's own loss experience or on credible aggregate experience from the same jurisdiction and line of business." The Bermuda benchmark did not meet this standard because it was not jurisdiction-specific. The regulator also noted that the index included claims from jurisdictions with different litigation environments, different securities laws, and different plaintiff bar incentives.

The carrier argued that the excess layer was so remote that the benchmark was a reasonable proxy. The regulator disagreed, pointing out that the layer had a nonzero expected loss and that the benchmark had not been adjusted for U.S. loss development patterns. The regulator required the carrier to file a supplemental actuarial memo that included a U.S.-only loss triangle, even if that triangle had limited data points.

The rule is not new—it has been part of NAIC guidance for decades. But in practice, many excess filings have been approved with limited scrutiny, especially for new market entrants who lack own experience. The rejection indicates stricter enforcement, particularly for D&O and other liability lines where jurisdiction-specific risk is material.

Another important regulatory dimension is the requirement for rate filings to be "not unfairly discriminatory." A rate that uses a global benchmark without adjustment may treat all policyholders equally but may still be discriminatory if it fails to reflect differences in risk. For example, a company with operations only in the U.S. would be charged the same rate as a multinational with exposure in Europe, even though their litigation risk differs. The regulator may view this as unfair discrimination, even if the rate is adequate on average. This adds another layer of scrutiny for filings that rely on non-U.S. benchmarks.

What the Actuarial Memo Revealed: Data Gaps and Assumptions

The supplemental actuarial memo that the carrier eventually filed revealed significant data gaps. The Bermuda index was based on a global database of D&O claims that included approximately 40% U.S. claims, 35% European claims, and 25% from other regions. The U.S. claims in the index were not separated by jurisdiction, so it was impossible to determine whether they reflected the same litigation climate as the state in question.

The memo showed that U.S. securities litigation frequency has been about 2–3 times higher than the global average over the past decade, driven by the Private Securities Litigation Reform Act and the active plaintiff bar. Severity also differs: U.S. class-action settlements average around $20–30 million, compared to $5–10 million in Europe. Using a global blended rate without adjustment would understate expected losses for a U.S.-only portfolio.

The carrier's 40% loading for jurisdictional risk was intended to bridge this gap, but the memo did not include a rigorous derivation. The loading was based on a judgmental comparison of U.S. to global loss ratios, not on a statistical analysis of the index's components. The regulator deemed the loading arbitrary and not data-driven, effectively rejecting the entire filing.

In the revised filing, the carrier constructed a U.S.-only loss triangle using data from a reinsurance broker's proprietary database, supplemented by publicly available Securities and Exchange Commission filings. The triangle had only 12 data points for the top $10 million layer, but the regulator accepted it because the data was jurisdiction-specific and the adjustments were clearly documented. The revised rate was approximately 12% higher than the original Bermuda-based rate.

The memo also highlighted a common actuarial technique: credibility weighting. With only 12 data points, the credibility assigned to the U.S.-only triangle was low—perhaps 10–20%—meaning the final rate was heavily influenced by a prior distribution. The carrier used a Bayesian approach, combining the sparse U.S. data with a broader industry loss ratio from U.S. admitted carriers. This approach satisfied the regulator because the prior was also U.S.-based, not global. The key lesson is that even thin domestic data, when combined with a domestic prior, is preferable to a global benchmark with a judgmental loading.

The Ripple Effect on D&O Pricing and Capacity

The rejection did not occur in isolation. Within weeks, several other carriers paused filings that used Bermuda-linked excess benchmarks. Lead underwriters for large D&O programs began shifting to domestic benchmarks, even if it meant using less credible data. The market for top-layer excess capacity tightened as reinsurers demanded U.S.-only loss experience before committing.

Pricing for top layers rose by an estimated 10–15% in the subsequent quarter, according to broker surveys. Some of this increase reflects the higher rate required by U.S.-only data; some reflects the reduced supply of capacity as carriers waited for regulatory clarity. A few Bermuda-based carriers withdrew from the U.S. excess D&O market entirely, citing the increased compliance burden.

The ripple effect extended to program structure. Brokers and risk managers who had built layered programs with a Bermuda excess layer had to reconsider the data sources for each layer. Some programs were restructured to use a single carrier for the entire tower, simplifying data consistency but reducing diversification. Others added a domestic excess layer below the Bermuda layer to absorb the risk that the Bermuda benchmark could not support.

The capacity tightening was most pronounced for smaller carriers with limited U.S. loss experience. Larger carriers with decades of domestic D&O data were less affected—they could already demonstrate rate adequacy using their own experience. The rejection thus reinforced the advantage of incumbents with credible data, a dynamic that may reduce competition in the long run.

For a related discussion of how jurisdictional differences affect policy wording and pricing, see this earlier article on a German D&O policy that paid a California board demand at a Frankfurt rate.

Another consequence is the increased cost of compliance for carriers. Preparing a U.S.-only loss triangle for an excess layer requires data that may not be readily available. Carriers must invest in data collection, cleaning, and analysis. For a new market entrant, this can add tens of thousands of dollars to the filing cost. Some carriers may decide that the U.S. excess D&O market is not worth the regulatory overhead, leading to reduced capacity and higher prices for buyers.

Practical Takeaways for Brokers and Risk Managers

For brokers and risk managers, the rejection offers several actionable lessons. First, ask whether the excess layers of a D&O program use domestic or offshore benchmarks. If a carrier relies on a Bermuda index or a London market survey, request the underlying data and ask how it has been adjusted for U.S. jurisdiction. The answer may reveal whether the rate is defensible under regulatory scrutiny.

Second, request actuarial justification for any benchmark loading. If the carrier applies a loading for jurisdictional risk, the loading should be derived from a transparent analysis—not a judgmental percentage. The memo should include a comparison of U.S. and global loss experience, with explicit adjustments for frequency, severity, and loss development.

Third, monitor state rate filing rejections for early signals. Rejections are public documents in most states, and they often reveal which benchmarks and assumptions regulators find problematic. A rejection in one state may foreshadow scrutiny in others, especially for multi-state programs.

Fourth, consider layered programs that use consistent data sources across all layers. If the primary and lower excess layers are priced using U.S. loss experience, the top excess layer should ideally use the same data, even if the data is sparse. Inconsistency invites regulatory challenge.

Finally, engage actuarial consultants early in the renewal cycle. A pre-filing review of the actuarial memo can identify data gaps and assumptions that regulators might question. The cost of an actuarial review is small compared to the cost of a rejected filing and a delayed market entry.

For a deeper look at how rate filings can hinge on specific assumptions, see this article on an adjuster who recalculated a hurricane loss from one roof fastener specification.

Brokers should also consider the timing of filings. If a carrier plans to enter a new state, the filing should be submitted well in advance of the desired effective date. A rejection can add months to the timeline, especially if the carrier must develop new data. Building a buffer of 6–9 months for the regulatory approval process is prudent, particularly for new entrants without an established track record in the state.

What the Next Filing Should Look Like

The next D&O filing that includes an excess layer should use domestic loss experience for all layers, even if the data is limited. Regulators have made clear that jurisdiction-specific data is preferred over global benchmarks, even when the domestic data has low credibility. The actuarial memo should include a sensitivity analysis that tests alternative benchmarks and shows the impact on the indicated rate.

Data sources should be clearly documented. If the carrier uses a reinsurance broker's database, the memo should describe the database's scope, the number of claims, and any adjustments made to align with the insured exposure. Publicly available data from the Securities and Exchange Commission or from litigation databases can supplement proprietary sources.

Adjustment factors should be derived from statistical analysis, not judgment. For example, if the carrier uses a global benchmark, it should estimate a jurisdiction-specific adjustment factor by comparing U.S. loss experience to global experience for the same line. The analysis should include confidence intervals and discuss the uncertainty around the adjustment.

The filing should be pre-filed with the regulator for feedback, especially if the carrier is new to the market or the program structure is unusual. A pre-filing review can identify issues before the formal submission, reducing the risk of rejection. Carriers should expect a 6–9 month lead time for approval, longer if the filing requires significant data development.

In the long run, the industry may need to develop a credible U.S.-only excess D&O benchmark. Several actuarial firms are working on this, but the data challenge is substantial. Until then, carriers will have to make do with thin data and transparent assumptions. The rejection of the Bermuda benchmark filing is a reminder that regulatory standards are not negotiable—and that a rate is only as good as the data that supports it.

One promising development is the creation of industry loss databases for excess layers. The NAIC has discussed a pilot program to collect excess-layer claims data from carriers, modeled on the existing medical malpractice closed claim database. If successful, such a database could provide credible U.S.-only benchmarks for excess D&O layers. However, the initiative is still in early stages, and participation is voluntary. Carriers should monitor these efforts and consider contributing data to help build a more robust foundation for rate filings.

Another approach is to use parametric triggers for excess layers. Instead of relying on loss experience, a parametric excess layer could be triggered by a predefined event, such as a stock price decline of a certain magnitude. The rate would be based on the probability of the trigger event, not on claims data. While parametric insurance is more common in property and catastrophe lines, it may gain traction in liability lines as a way to bypass the data problem. Regulators would still need to approve the trigger mechanism, but the rate adequacy argument would be different—based on financial modeling rather than loss triangles.

This article is for informational purposes only and does not constitute professional advice. Readers should consult qualified actuaries and legal counsel for specific rate filing and regulatory compliance matters.