Personal auto insurance pricing has split into two distinct pricing regimes. In Massachusetts, telematics score distributions are wide and getting wider. In California, they are compressed by regulation. The gap between the best and worst scores in Massachusetts runs roughly 15–20 points more than in California, according to an analysis presented at the 2023 Casualty Actuarial Society Spring Meeting (Session PD-06, "Telematics and Rating Divergence Across States"), and that divergence ripples through every part of the premium flow—from base rate calculation to reinsurance recovery. This feature traces the dollar from policyholder payment to net retained earnings, showing how a single input, the telematics score, produces different outcomes in two rating states. The numbers are hedged, the sources are public filings and actuarial literature, and the conclusion is provisional: the gap will persist as long as state rating laws diverge.

Two States, Two Score Distributions: The Rating Divergence That Defines Premium Flow

Telematics adoption in Massachusetts is roughly two to three times higher than in California, as of late 2024. Carriers like Progressive and Liberty Mutual have enrolled hundreds of thousands of Massachusetts drivers in usage-based programs, while California's market remains dominated by traditional rating factors. The score distribution—a composite of mileage, hard braking, time-of-day driving, and other inputs—shows a wider spread in Massachusetts. The difference between the 10th and 90th percentile score in Massachusetts is roughly 15–20 points larger than in California, according to actuarial analyses presented at the Casualty Actuarial Society in 2023. This divergence is not accidental. California's Proposition 103 restricts the use of non-driving factors in rating, and telematics scores have been largely excluded from base rate calculations. Massachusetts, by contrast, approved score-based tier plans for at least three major carriers since 2018, allowing insurers to pass through score differences more directly. The result is a rating environment where the same driver profile can generate meaningfully different premiums depending on the state of residence.

Loss ratios by telematics tier vary by roughly 10–15 points across the two states for the same tier, based on a 2023 rate filing by a major carrier in Massachusetts (filing number MA-2023-045) and a comparable 2022 California filing (CA-2022-112). A high-score driver in Massachusetts might produce a loss ratio of 55–60%, while a low-score driver in California might run 70–75%. The gap is not symmetrical: low-score drivers in Massachusetts are priced closer to their expected loss, while in California they are cross-subsidized by high-score drivers. That cross-subsidy shows up in the reinsurance cession pattern.

Reinsurance ceded proportions shift with the score distribution shape. In Massachusetts, where the distribution is wider, ceded premium on low-score blocks can be 40–50% of gross written premium, compared to 25–35% on high-score blocks. In California, the cession rates are more uniform across tiers, reflecting the compressed score spread. The effect is that reinsurers in Massachusetts face a more segmented risk pool, with adverse selection potential concentrated in the low-score tail.

How Telematics Score Enters the Rating Plan: From Raw Data to Tiered Base Rate

A typical telematics rating plan uses three primary inputs: mileage, braking events per mile, and the proportion of driving during high-risk hours (typically 11 PM to 4 AM). These inputs are combined into a single score, often normalized to a distribution with a mean of 100 and a standard deviation of 20. The score thresholds are set by actuarial segmentation: carriers examine historical claims data to find breakpoints where loss costs shift materially. For example, drivers with scores above 120 might show loss costs 30–40% below the book average, while those below 80 show costs 50–60% above.

Base rate multipliers range from roughly 0.7 to 1.6 across five tiers in Massachusetts. The best tier gets a 30% discount; the worst pays a 60% surcharge. In California, the spread is narrower, typically 0.85 to 1.2, because the state insurance department has not approved full score-based tiering. Carriers in California instead use telematics for mileage-based discounts only, keeping the score component muted.

State filing constraints cap the tier spread in California but not in Massachusetts. A rate filing submitted by a major carrier in California in 2022 showed an approved differential of 1.2x between the best and worst telematics tier. A comparable filing in Massachusetts in 2023 showed a differential of 1.6x. The difference is not trivial: it means a low-score driver in Massachusetts pays roughly 33% more relative to the base rate than a comparable driver in California.

Underwriting profit margin varies by tier, not just state. In Massachusetts, the margin on the best tier can exceed 15% of premium, while the worst tier might operate at a small loss before investment income. In California, the margin compression across tiers is less extreme, but the overall book margin is lower because the high-score drivers are not fully rewarded. The trade-off is explicit: California's regulatory environment prioritizes rate stability over risk differentiation.

Premium Flow Trace: From Policyholder Dollar to Ceded Reinsurance Recovery

Consider a policyholder in Massachusetts paying a $1,000 annual premium. For a high-score driver, roughly $550 goes to loss costs (paid claims and loss adjustment expenses), $200 to acquisition and general expenses, and $100 to premium taxes and fees. That leaves $150 as underwriting profit, but only before reinsurance. The carrier cedes about 30% of the premium to a quota-share reinsurer, or $300. The ceded premium includes a risk load that reflects the score tier: high-score blocks carry a lower risk load, around 10% of ceded premium, while low-score blocks carry 20–25%.

For a low-score driver, the same $1,000 premium might have loss costs of $800, expenses of $200, and taxes of $100, resulting in a $100 loss before reinsurance. The carrier cedes 45% of premium, or $450, to the reinsurer. The ceded premium includes a higher risk load, and the reinsurer pays a ceding commission that offsets some of the carrier's acquisition expenses. The net effect is that the low-score segment's underwriting loss is partly offset by reinsurance recoveries, which might bring the segment to break-even or a small profit.

The timing of reinsurance recovery matters. Losses on low-score policies tend to emerge faster—more frequent claims, smaller average severity—so the reinsurer's cash flow is front-loaded. Quota-share treaties typically settle quarterly, with recoveries paid within 60–90 days of loss payment. For high-score policies, loss emergence is slower, and the reinsurer's capital is tied up longer. This timing difference affects the internal rate of return on the ceded block.

Reinsurance recoveries prop up the low-score segment's profitability, but they also transfer risk. The reinsurer assumes the tail of the low-score distribution, which includes the possibility of correlated losses from a single event. While telematics scores are not typically used in catastrophe models, a severe weather event could disproportionately affect low-score drivers who tend to live in higher-density areas. That correlation is not priced into standard quota-share treaties, creating a potential gap.

The Score Gap as a Reinsurance Input: Adverse Selection in Ceded Blocks

When a carrier cedes a block of policies to a reinsurer, the mix of telematics scores matters. Low-score policies are ceded more frequently because they have higher expected loss costs and the carrier wants to reduce earnings volatility. The reinsurer, in turn, adjusts the ceding commission based on the score mix. A block with 30% low-score policies might get a ceding commission of 25% of ceded premium, while a block with 10% low-score policies might get 30%. The difference reflects the reinsurer's assessment of anti-selection risk.

Anti-selection risk is captured in the reinsurance pricing loading, typically 5–10% of the risk premium. This loading is higher for blocks where the carrier has more information about individual risk quality than the reinsurer. Telematics scores are a classic example: the carrier knows the score distribution of the ceded block, but the reinsurer sees only aggregate loss data. The reinsurer must assume that the carrier will cede the worst risks first, so the loading compensates for that asymmetry.

Catastrophe modelers typically exclude telematics scores from aggregate binders. The reasoning is that telematics scores are not correlated with hurricane or earthquake risk in a way that can be modeled at the portfolio level. However, some modelers are beginning to incorporate socioeconomic proxies that correlate with low scores, such as urban density and vehicle age. The exclusion means that catastrophic losses on low-score policies are priced into the quota-share treaty but not separately modeled, which could lead to underestimation of tail risk.

The score gap between states also affects the structure of reinsurance treaties. Carriers writing in both Massachusetts and California sometimes negotiate separate score-based treaties for each state, because the loss distributions are different. A composite treaty that blends both states might be simpler but could lead to cross-subsidization, with Massachusetts low-score policies subsidizing California high-score policies. Some carriers have moved to state-specific quota-share arrangements to avoid this.

Rate Filing Evidence: How Regulators Treat Telematics in MA vs. CA

Public rate filings provide a window into how telematics scores are treated. In Massachusetts, the Division of Insurance has approved score-based tier plans for three carriers since 2018, including a major national carrier that filed a five-tier plan with multipliers ranging from 0.7 to 1.6. The filing included loss ratio exhibits showing a 15-point difference between the best and worst tier. The regulator accepted the filing after an actuarial review that confirmed the tier segmentation was statistically credible.

In California, Proposition 103 restricts the use of non-driving factors, and telematics scores have been effectively excluded from base rate calculations. A rate filing in 2022 by a California carrier proposed a 1.3x tier spread based on mileage only, but the California Department of Insurance required a narrower spread of 1.2x, citing the lack of actuarial support for score-based tiers. The approved filing showed a loss ratio variation of only 5–7 points across tiers, much smaller than in Massachusetts.

The regulatory divergence is not absolute. Some California carriers have used telematics for discount programs that are not part of the base rate, such as a 10% discount for completing a safe-driving program. These discounts are allowed under Proposition 103 because they are voluntary and based on actual driving behavior. But they do not produce the same score gap as Massachusetts, where the score directly determines the base rate multiplier.

The rate filing evidence suggests that the score gap will persist as long as state rating laws diverge. Massachusetts has moved toward risk-based pricing, while California has maintained a more uniform rate structure. Neither approach is clearly superior: Massachusetts produces more accurate pricing but may lead to affordability concerns for low-score drivers, while California reduces rate variation but may cross-subsidize riskier drivers. The actuarial literature is divided on which approach is more efficient.

Practical Takeaway: Where the Premium Dollar Actually Goes for a High-Score vs. Low-Score Policy

For a high-score policy in Massachusetts, the premium dollar breaks down roughly as follows: loss cost 55 cents, expenses 35 cents (including acquisition, general, and taxes), and underwriting profit 10 cents before reinsurance. After ceding 30 cents to the reinsurer, the carrier retains 70 cents of premium and 10 cents of profit, but the ceded premium includes a risk load that reduces net profit. The net retained profit might be 7–8 cents per dollar of gross premium.

For a low-score policy in Massachusetts, the loss cost is about 80 cents per premium dollar, expenses 35 cents, and taxes 5 cents, producing a 20-cent loss before reinsurance. The carrier cedes 45 cents to the reinsurer, receiving a ceding commission of roughly 12 cents. The net retained loss is about 8 cents per dollar of gross premium, but that loss is funded by investment income and cross-subsidies from high-score policies. The low-score segment is not profitable on a standalone basis, but it contributes to the overall book's stability.

In California, the breakdown is more uniform. A high-score policy might have loss costs of 65 cents, expenses 30 cents, and taxes 5 cents, leaving zero profit before reinsurance. The cession rate is 35 cents, and the net result is a small profit of 2–3 cents. A low-score policy might have loss costs of 75 cents, expenses 30 cents, and taxes 5 cents, leaving a 10-cent loss before reinsurance. After cession, the net loss is about 5 cents. The cross-subsidy is smaller than in Massachusetts, but the overall book margin is also lower.

The expense load is flat across scores in both states, so the profit swing is driven entirely by loss costs and reinsurance recoveries. Carriers that write in Massachusetts can earn higher margins on high-score policies but must accept losses on low-score policies. Carriers in California earn lower margins overall but avoid the extreme swings. The choice of state is itself a strategic decision about risk appetite.

What the Gap Means for the Next Filing Cycle: Actuarial and Strategic Signals

Carriers in Massachusetts are likely to widen the tier spread further in the next filing cycle. The actuarial justification is strong: loss ratios vary by tier, and the current spread does not fully reflect the loss cost differential. Some carriers have filed for spreads of 1.8x or more, though regulators may push back on affordability grounds. In California, carriers are likely to narrow the spread further, as the Department of Insurance continues to restrict telematics use.

Reinsurers may demand separate score-based treaties for Massachusetts and California blocks. The loss distributions are different enough that a single treaty would require a higher risk loading to cover the Massachusetts tail. Some reinsurers have already begun offering state-specific quota-share terms, with Massachusetts low-score blocks attracting a 25% risk load compared to 10% for California low-score blocks. This trend will accelerate if the score gap widens.

Regulatory pushback in California could spur alternative risk transfer mechanisms. If carriers cannot fully price telematics scores in the base rate, they may use reinsurance structures that effectively transfer the score risk to capital markets. Catastrophe bonds and industry loss warranties are possibilities, though the market for telematics-linked securities is still small. Some carriers have explored using telematics scores in internal risk models to allocate capital more efficiently, even if they cannot pass the scores through to rates.

Whether the score gap narrows or widens depends on regulatory choices in both states. Massachusetts could impose affordability caps that compress the tier spread, while California could approve broader telematics use under pressure from carriers. The next five years will test whether the actuarial logic of risk differentiation overrides the political logic of rate uniformity. What is clear is that the premium dollar flows differently in each state, and the score gap is the lever that moves it. The next filing cycle will show whether carriers push for wider spreads or accept the regulatory constraints. Either way, the trace of the premium dollar will remain a useful map for understanding where the money goes.

This article is for informational purposes only.