Clearspeed, a voice-based risk-assessment vendor that sells fraud-detection tools to insurers, released a new analysis through PR Newswire in mid-August 2026 estimating that fraud and inaccurate information cost the U.S. personal auto and homeowners insurance market an estimated $45.3 billion annually. The report was authored by Kim Garland, who is described as a veteran insurance executive and Clearspeed advisor, and as former President of State Auto Insurance and former Chief Product Officer of AIG’s Global P&C Consumer Division — not an independent or neutral observer. Clearspeed’s analysis does not disclose its underlying methodology: the release names no data sample, no data source, and no time period for the figures it reports.
Where Clearspeed’s estimate comes from
The headline number appears in a Clearspeed-commissioned paper titled ‘The next multi-billion dollar variable in the insurance industry.’ distributed alongside a companion piece titled ‘Homeowners: The P&C insurance market most likely to fail over the next five to ten years.’ Both documents were pushed out the same week through the company’s own newswire distribution rather than through a peer-reviewed journal, a rating agency, or a government statistical office. Readers looking for the sampling method, the claims data behind it, or the calculation model will not find one in the public version of the release. That absence matters because the figure has already begun circulating in industry conversation as though it were independently verified market data.
Splitting the number between auto and homeowners
Clearspeed’s total is broken down by line of business, with approximately $31.6 billion annually in personal auto insurance and $13.7 billion annually in homeowners insurance. The homeowners share arrives at a moment when carriers are already tightening underwriting in that line; pricing pressure already building in homeowners coverage has been visible in state-level data well before this analysis appeared. Clearspeed frames its estimate as covering both outright fraud and, separately, information that policyholders submit inaccurately, whether deliberately or not — a broader category than fraud alone, though the release does not quantify how the two components split within the total, nor does it explain why homeowners, a smaller line by premium volume than auto, accounts for such a large share of the estimated loss.
Trustworthiness as underwriting’s next variable
The analysis’s central argument goes beyond the dollar figure. Garland writes that trustworthiness may represent the first new transformational underwriting variable in personal lines insurance since the introduction of credit scoring. In his framing, the U.S. insurance system has evolved into a low-trust system that cannot run indefinitely, as the costs of a few overwhelm the many, and rate increases alone are not a sustainable response. He goes on to argue that if insurers could practically separate trustworthy from non-trustworthy actors at each decision point, this could reverse a decades-long trend, adding that as more large carriers adopt such tools, the question becomes how long competitors can afford to wait. Clearspeed also states that its insurer clients experience, on average, a 40% increase in immediate settlements, a 50% reduction in claims-handling time, and a 36% increase in preventive fraud savings — figures the release attributes to the company’s own client base rather than to an independent study, without specifying how many clients or what period those averages cover.
How the estimate compares with other fraud figures
Clearspeed’s $45.3 billion figure is not the first attempt to size fraud’s cost to personal lines insurers, and it is worth reading alongside older, differently constructed estimates rather than as a replacement for them. According to a 2017 study by Verisk cited by the Insurance Information Institute, auto insurers lose at least $29 billion a year to fraud and misrepresentation — an earlier estimate built on a different methodology than Clearspeed’s auto-line figure, and one that predates the current analysis by nearly a decade. The same body of research, summarized separately by the Insurance Information Institute, also finds that as much as 14 percent of all personal auto premiums can be attributed to the cost of covering premium leakage, a category spanning both fraud and misrepresented application information. None of these figures were produced using the same methodology or the same time period, so they should be read as separate data points that bracket a similar problem rather than as numbers that confirm one another. Fraud pressures building in other markets too suggest the underlying concern extends well beyond the United States, even though every dollar estimate in this article is U.S.-specific.
Who is behind the analysis
Context on authorship matters here. Kim Garland is a Fellow of the Casualty Actuarial Society with more than three decades of executive leadership, and his prior roles at State Auto Insurance and AIG give him direct industry standing. But he is also, by the release’s own description, a paid advisor to Clearspeed, and Clearspeed describes itself as founded in 2016 and ‘the global leader in voice-based risk assessment,’ with proprietary technology that assesses risk based on vocal characteristics said to be universal to all humans — the same technology it sells to the insurers this analysis argues should adopt trustworthiness scoring. That commercial relationship does not make the underlying numbers wrong, but it does mean the $45.3 billion figure comes from a party with a direct interest in insurers accepting its premise, delivered through a press release rather than an audited study. Readers evaluating the claim should treat it as a vendor-commissioned estimate rather than as independently verified market data, much as disputes over claims disclosure practices elsewhere in the industry often turn on which party controls the disclosure. The episode also echoes broader questions about fraud risks tied to newer distribution channels, where the source of a claim shapes how much weight it deserves.
