Amazon Tax Dispute Triggers Rare $300M-$400M Transaction Liability Insurance Loss

Executive Summary

A decade-old sales tax dispute between Amazon and the state of South Carolina reached its final chapter on March 18, 2026, when the state Supreme Court ruled 3-2 that Amazon was legally required to collect and remit sales taxes on third-party marketplace transactions. The immediate assessment at stake was $12.5 million. But the ruling's logic extends to an additional $277 million in pending assessments covering 2016 through 2019. When combined with interest, penalties, defense costs, and contractual gross-up provisions, the total exposure is expected to trigger a rare full-limit payout of $300 million to $400 million under a transactional liability insurance policy placed across multiple specialty carriers.

Media reports have identified QBE, Chubb, DUAL, and Euclid Transactional among the insurers on the risk. This represents one of the largest known single-risk losses in the transactional liability insurance sector, a specialty line that has attracted significant capital in recent years on the strength of favorable loss experience. For actuaries working in specialty lines reserving, pricing, or enterprise risk management, this event sits at the intersection of evolving tax law, digital economy platform liability, and the rapid maturation of a coverage line now confronting its first generation of severe losses.

The Tax Dispute: From Exemption to Assessment

The origins of this case reach back to 2011, when Amazon was negotiating plans to build distribution centers in South Carolina. The state legislature passed Act 32 of 2011, granting Amazon a five-year exemption from collecting sales taxes in exchange for a commitment to create at least 2,000 full-time jobs and invest $125 million. Amazon subsequently built its first two distribution facilities in the state, in West Columbia and Spartanburg.

When the exemption expired on December 31, 2015, Amazon began collecting and remitting South Carolina sales taxes on products sold directly by Amazon and its affiliates. It did not, however, collect taxes on the roughly 50 percent of platform sales volume generated by approximately 2.5 million independent third-party merchants who listed and sold products through the Amazon Marketplace.

The South Carolina Department of Revenue (DOR) audited the first quarter of 2016 and assessed $12,490,502.15 in unpaid taxes, penalties, and interest. Amazon contested the assessment, arguing that it was not "engaged ... in the business of selling tangible personal property at retail" under S.C. Code Ann. Section 12-36-910(A). The company's position was that third-party merchants were the sellers, that Amazon's role was limited to providing a platform and ancillary services, and that the statute was at minimum ambiguous and should therefore be construed in the taxpayer's favor.

The Platform Control Question

The legal battle hinged on a deceptively simple question: when a consumer searches for a product on Amazon.com, selects it, pays for it, receives it via Amazon logistics, and processes any return through Amazon's systems, who is the "seller"?

Amazon pointed to its Business Solutions Agreement, which designated third-party merchants as the sellers of record. But the state argued that this contractual label did not reflect economic reality. Under the same agreement, Amazon exercised extensive control over virtually every aspect of the transaction: product listing standards, pricing rules, customer-facing communications, payment processing through Amazon Payments, shipping and fulfillment coordination, returns handling, and fund disbursement to merchants only after deducting Amazon's fees.

The court found that this level of operational control made Amazon "integral" to every third-party transaction. Without Amazon's active involvement, these sales could not have occurred on the platform. The distinction between being a "service provider" and being "engaged in the business of selling" collapsed under the weight of the platform's actual role in facilitating, controlling, and profiting from these transactions.

The Ruling and the Dissent

The South Carolina Supreme Court issued its opinion on March 18, 2026, in Case No. 28319, affirming the lower courts' decisions.

Justice John Cannon Few wrote for the majority, joined by Justices D. Garrison Hill and George C. James Jr. The opinion held that the Sales and Use Tax Act's language was "unambiguous" as applied to Amazon Services. The majority applied ordinary dictionary definitions of "engaged" (meaning "involved in activity" or "employ or involve oneself") and found that Amazon's comprehensive control over third-party transactions plainly met this threshold. Because the statute was unambiguous, there was no occasion to invoke the rule of strict construction that favors taxpayers in cases of statutory doubt. The court also rejected Amazon's argument that applying the 2016 law to its operations violated due process, finding that Amazon had sufficient notice of the state's tax collection obligations.

Chief Justice John Kittredge and Justice Courtney Clyburn Pope dissented, arguing that the statute's application to marketplace facilitators was genuinely ambiguous and should have been construed against the government.

The ruling upheld two previous decisions: the South Carolina Administrative Law Court's 2019 ruling (following a three-day evidentiary hearing) and the Court of Appeals' 2024 affirmance (442 S.C. 313).

From $12.5 Million to $300 Million: How the Exposure Escalated

The Supreme Court case addressed only the $12.5 million Q1 2016 assessment, which Amazon had already paid as a condition of pursuing its appeal. But the ruling's reasoning applies with equal force to the entire period during which Amazon collected no third-party sales taxes in South Carolina.

Amazon did not begin collecting taxes on third-party merchant sales until April 30, 2019, after the state legislature passed marketplace facilitator legislation. The Department of Revenue conducted two additional audits covering April 2016 through 2019 and issued assessments totaling approximately $277 million in back taxes and interest. Those cases were stayed at the Administrative Law Court pending the Supreme Court's resolution of the underlying legal question.

With the legal principle now definitively resolved against Amazon, the path to collection on those additional assessments has been substantially cleared. A Revenue Department spokesman confirmed on the day of the ruling that Amazon had not yet paid the $277 million.

When the total tax liability is combined with accruing interest, potential additional penalties, defense costs incurred over a decade of litigation, and the gross-up provisions standard in tax liability insurance policies (which compensate the insured for income taxes owed on the receipt of insurance proceeds), specialty market sources estimate the total insured loss at between $300 million and $400 million.

The Insurance Story: A Rare Full-Limit TL Loss

This is where the dispute shifts from a state tax case into a significant event for the specialty insurance market.

Amazon evidently purchased transactional liability insurance, specifically a standalone contingent tax liability policy, to hedge the risk that its pre-Wayfair sales tax position in South Carolina would fail. This is precisely the kind of exposure that tax liability insurance is designed to address: a known but uncertain tax position where a taxpayer has taken a defensible stance that nonetheless carries meaningful risk of being overturned.

What Is Tax Liability Insurance?

Tax liability insurance is a specialized product within the broader transactional liability (TL) market, which also includes representations and warranties insurance (R&W, used in M&A transactions) and contingent risk insurance. Tax insurance is designed to make the insured whole in the event that a taxing authority successfully challenges the insured's tax position. Coverage typically includes the additional taxes owed, associated interest and penalties, the costs of defending the position, and a "gross-up" to compensate for income taxes assessed on the receipt of the insurance proceeds.

The product has grown significantly over the past decade as corporations, private equity sponsors, and deal parties have increasingly used insurance to ring-fence known tax risks. It functions somewhat like obtaining a private letter ruling from the IRS, but it can typically be purchased within weeks rather than the months or years a formal ruling requires. Premiums are priced primarily on the basis of legal opinion about the probability that the insured tax position will be sustained.

The Carriers on the Risk

Insurance Insider US reported in September 2024 that the transactional liability sector was bracing for a potential full-limit loss of up to $400 million from the Amazon-South Carolina dispute. According to that reporting, the tax litigation policy was brokered by CAC Specialty (the specialty-insurance brokerage arm of CAC Group, which has announced a planned merger with The Baldwin Group expected to close in 2026). The policy was placed across both U.S. and London specialty markets.

Media reports have identified four major specialty carriers among those exposed:

QBE. One of the largest global insurers, QBE has been actively expanding its transactional liability operations. In March 2026 alone, QBE appointed its first Asia-based transactional liability underwriter and announced an algorithmic underwriting partnership with Aurora. The global head of QBE's transactional liability practice is Toria Lessman. QBE posted statutory net profit of $2.157 billion for FY2025, with a group combined operating ratio of 91.9 percent. However, the specialty business posted a combined operating ratio above 100 percent, with prior-year adverse development in transaction liability specifically cited as a contributing factor.

Chubb. Chubb offers tax liability, representations and warranties, and contingent risk products through its global transactional risk team. It is one of the largest writers of transactional risk insurance worldwide. Chubb reported approximately $700 million in adverse loss development for the other liability (occurrence) line during 2025.

DUAL. DUAL North America (part of Howden Group) specializes in representations and warranties, tax liability, and contingent liability insurance, with capacity to underwrite transactions with limits exceeding $30 million either as lead or participant. Its transactional risk team maintains appetite across all enterprise value ranges and a wide spectrum of industry sectors.

Euclid Transactional. A leading independent underwriter of transactional insurance products, Euclid writes on CNA/Columbia Casualty paper (surplus lines) and has insured over $8 trillion in total deal value across more than 9,100 policies since its founding in 2016. Its products span representations and warranties, tax liability, and contingent liability coverage.

Why This Loss Is Significant

The loss is described by specialty market observers as "rare" in the TL sector, where full-limit payouts on individual tax insurance policies of this magnitude are uncommon. Tax insurance underwriting typically involves detailed legal analysis of the insured position's merits before a policy is bound. The fact that Amazon's position was ultimately rejected by three successive South Carolina tribunals raises questions about the initial underwriting assessment of the risk, though it is worth noting that the Supreme Court itself was divided 3-2, that major business organizations including the U.S. Chamber of Commerce filed amicus briefs supporting Amazon's position, and that the underlying legal question (whether pre-Wayfair statutes could impose marketplace facilitator obligations) was genuinely novel.

What This Means for the TL Market

From tracking specialty market developments over the past two years, several features of this loss event carry broader implications.

Scale relative to market norms. A $300-$400 million full-limit loss on a single tax insurance placement is extraordinary. Tax insurance losses tend to be partial recoveries or negotiated settlements rather than full-limit payouts triggered by a definitive adverse judicial ruling. This loss will land as a significant data point for any actuarial analysis of severity distributions in the tax insurance line.

Reserving timeline complexity. This loss has developed over an unusually long arc. The initial assessment was December 2016. Insurance Insider reported the potential $400 million exposure in September 2024. The Supreme Court ruling crystallized the loss on March 18, 2026. For carriers on the risk, the reserving question is whether this was fully case-reserved at initial notification, incrementally built through the appellate process, or whether the ruling triggers material reserve strengthening in Q1 2026 results. QBE's disclosure of prior-year adverse development in transaction liability for FY2025 suggests the loss was at least partially recognized before the ruling, but the gap between "partially reserved" and "full-limit payout" can still be significant.

Concentration risk in syndicated placements. When a single large policy is syndicated across multiple carriers and markets, a single adverse judicial outcome creates simultaneous losses for all participants. This is structurally analogous to how a single catastrophe event generates correlated losses across property reinsurance panels. Actuaries should consider whether their transactional liability portfolio models adequately capture the correlation inherent in syndicated placements, particularly where multiple policies in the book may share exposure to the same legal theory or regulatory trend.

Pre-Wayfair tail risk. This ruling addresses the gap period before both the U.S. Supreme Court's 2018 Wayfair decision (which eliminated the physical-presence requirement from Quill) and the subsequent wave of state marketplace facilitator laws. The Sales Tax Institute noted following the original 2019 ALJ ruling that "other marketplaces could also be found to be responsible for the collection of taxes for periods prior to the enactment of marketplace facilitator legislation." If Amazon or other marketplace facilitators purchased tax insurance for analogous pre-Wayfair positions in other states, additional claims may follow.

Market response. The likely response is not a wholesale retreat from tax insurance underwriting but a recalibration: higher attachment points for very large limits on novel positions, more rigorous independent legal review, tighter policy language around covered tax periods, and elevated pricing for positions involving regulatory frameworks in active transition.

Actuarial Implications

For actuaries working in specialty lines, this event surfaces several considerations.

Pricing frameworks need scenario-based stress testing. Tax liability insurance pricing relies on legal opinion about the probability that a position will be sustained. There is no frequency-severity model derived from thousands of historical claims. A position that legal counsel rates at 75 or 80 percent probability of success still carries a 20-25 percent probability of a full-limit loss. When limits are in the hundreds of millions, the expected loss contribution from that tail scenario is substantial. Actuarial pricing models should incorporate explicit scenario analysis alongside legal opinion.

Legal-theory correlation is underappreciated. Multiple policies in a TL book may share exposure to the same underlying legal theory. A single court ruling can trigger claims across the portfolio simultaneously. This is not independent risk. Actuaries building aggregate loss models need to identify and model these legal-theory clusters, much as property catastrophe modelers identify geographic concentration.

Development patterns are driven by litigation, not reporting lags. The nearly ten-year duration from initial assessment to Supreme Court ruling illustrates how long-tailed certain specialty exposures can be. Standard three- or five-year development triangles may be insufficient. Actuaries setting IBNR reserves should build extended development patterns and conduct case-level reviews as litigation milestones are reached.

Enterprise risk and capital implications. For carriers with meaningful TL portfolios, a $300-$400 million loss from a single risk is a potentially material event. This type of scenario belongs in internal model stress tests and in the Own Risk and Solvency Assessment (ORSA) process.

Looking Ahead

Several questions remain open. Amazon may seek review from the U.S. Supreme Court, though certiorari for a state-law tax interpretation case is historically rare. The two pending DOR assessments totaling $277 million will now advance through the Administrative Law Court, though the central legal question has been resolved.

For the transactional liability insurance market, this loss will be felt not just in the financial results of the carriers involved but in the underwriting discipline applied to the next generation of tax insurance submissions. The event provides a concrete, large-scale case study that will inform actuarial models, underwriting guidelines, and broker submissions for years to come.

For actuaries, the Amazon-South Carolina saga is a powerful reminder that specialty lines carry distinctive risk characteristics that resist easy analogies to traditional casualty. Tax insurance losses are binary, driven by judicial outcomes rather than gradual claim development. They can be very large on a per-risk basis. And they arise from an area of law subject to ongoing evolution. Sound reserving and pricing in this space requires not just actuarial technique but a working understanding of the legal and regulatory currents that generate the underlying exposures.