Texas Supreme Court Says Amazon Isn’t the Seller of Marketplace Items It Fulfills–Amazon v McMillan

The Fifth Circuit certified the following question to the Texas Supreme Court:

Under Texas products-liability law, is Amazon a “seller” of third-party products sold on Amazon’s website when Amazon does not hold title to the product but controls the process of the transaction and delivery through Amazon’s Fulfillment by Amazon program?

On a 5-2 vote, the Texas Supreme Court says no.

The Majority Opinion

The majority links the definition of “seller” to whomever transfers the title:

In cases where an ordinary sale takes place, we have never held that a seller can be anyone other than the person or entity who relinquishes title. And nothing in Chapter 82 signals that the Legislature intended to expand seller liability beyond that available under the common law…. Chapter 82 does not expand the pool of potentially liable non-manufacturing sellers beyond those recognized at common law; it reduces that pool….When the facts involve a sale in the ordinary sense, none of our cases have suggested that anyone other than a person or entity who has relinquished title in the chain can be a “seller.”…when a non-sale commercial transaction occurs, none of our cases have suggested that anyone other than those who have at least held title in the chain can be a seller

The majority draws some comfort from the fact that “many courts have concluded that Amazon is not a seller when it does not hold or relinquish title to the product.” Cites to State Farm v. Amazon (9th Cir.), Erie Insurance v. Amazon (4th Cir.), and Eberhart v. Amazon (SDNY), with additional support from Fox v. Amazon and Garber v. Amazon.

The majority concludes: “when a product-related injury arises from a transaction involving a sale, sellers are those who have relinquished title to the allegedly defective product at some point in the chain of distribution.”

As applied to Amazon, this becomes an easy case. Amazon didn’t hold the title; the vendor’s items were “entrusted” to it. Amazon’s possession of the items doesn’t change the analysis; “none of Amazon’s upstream actions—such as web hosting, advertising, marketing, warehousing, payment processing, and shipping—constituted a sale under Texas law.” The court summarizes:

Amazon has developed a business model that is not title dependent. And though Amazon’s business model gives the company a significant amount of control over the process of the transaction and the delivery of the product, this control does not transform an otherwise non-title-holding sales facilitator into a “seller.”

The Dissenting Opinion

The dissent has a different interpretation of the statutory words “distributing” and “placing”: “Amazon.com completely ‘controls’ the entire ‘process of the transaction’ as well as the ‘delivery’ of the product to the ultimate consumer.” The dissent says the majority’s bright-line title-dependent test is “helpful” but jurisprudentially novel in Texas. So while the dissent agrees with the majority on the policy question, the dissent disagrees as a matter of statutory construction.

(I’m always fascinated when judges display their textualism proudly but reach conflicting results. As I recently said, there isn’t one definitive way to do textualist interpretations, which means that even hardcore textualists can import their own normative assumptions into their methodology).


As the dissent says, focusing on transfer of title “draws a bright and logical line that can promote certainty and predictability in the law.” Unfortunately, the law on this topic has become anything but predictable or certain.

For example, Texas’ law collides with California’s, which held first that Amazon was strictly liable for items it fulfills (Bolger) and then doubled-down to hold Amazon liable for items it didn’t fulfill and never physically possessed (Loomis). (Strangely, neither the majority nor the dissent engage with that precedent). It would be tempting to treat California courts as the outlier, but some other courts around the country–notably the Oberdorf opinion, until it quietly fizzled out–have been on California’s wavelength.

So where does this leave us? Unless Amazon fixes its California problem, winning in other states is nice but incomplete. Recall that Amazon was ready to take a statutory deal in California that would have imposed strict liability on it, so long as its competitors were equally treated, so it’s a little hard to be too sympathetic to the legal plight Amazon finds itself in. Ultimately, unless Amazon magically finds a way to run the table in the courts, a statutory resolution is the only stable end game. As I’ve predicted before, any final resolution will almost certainly include Amazon restricting or shutting down its marketplace and affirmatively acting as the retailer for most or all of the items it sells–a move that will likely exacerbate, not alleviate, Amazon’s market power and hurt many small businesses currently thriving in its marketplace.

Case citation: Amazon.com, Inc. v. McMillan, No. 20-0979 (Tex. Supreme Ct. June 25, 2021). Majority opinion. Dissenting opinion.

Related Blog Posts: