This coming spring, the Supreme Court will hear arguments in the United States v. Microsoft – a case that will determine the authority of U.S. law enforcement to compel, via a warrant, US-based companies to turn over data held outside the United States. Over at Lawfare, Orin Kerr posits that Microsoft and the government—as well as the numerous lower court judges that have weighed in—have missed the core issue in the case. According to Kerr, the key is the All Writs Act; the parties and lower court judges have, in contrast, all focused on the Stored Communications Act. According to Kerr, only the All Writs Act gives the Supreme Court the necessary latitude to craft the kind of nuanced response that is needed.
This is a more detailed reprise of a claim that Kerr made some two year ago. I disagreed then (see our back and forth here). And I disagree now.
A bit of background: The dispute centers around a four-year old warrant that the government issued on Microsoft seeking emails associated with a particular account. Microsoft refused, on the grounds that the data was stored in Ireland, that the warrant authority is territorially-limited, and that therefore the compulsion order constituted an impermissible extraterritorial application of the government’s warrant authority. The government, by contrast, argued that since the warrant was served on Microsoft, and Microsoft could access the data from its offices in Redmond, Washington, without ever setting foot in Ireland, this was a territorial exercise of authority.
Both Microsoft and the government (including in their Supreme Court filing just yesterday) – as well every judge that has looked at the issue — have concluded that the key to the case is the SCA. Enacted in 1986, the SCA both prohibits service providers such as Microsoft from voluntarily disclosing the content of customer data, absent the application of specifically identified exceptions, and specifies the procedures that the government must comply with to compel production of such data. Parties have agreed that the statute does not have extraterritorial reach, but have disagreed about what constitutes the “focus” of the statute and thus whether the compelled production order is a territorial or extraterritorial action.
Kerr makes two key points. First, he says that the SCA doesn’t actually provide the authority to compel production, but merely refers back to pre-existing authority—namely Rule 41 warrants, coupled with the All Writs Act. Second, Kerr argues that All Writs Act provides the Court the needed flexibility to craft a thoughtful and appropriate response to the case.
I’d like to agree with Kerr. I am, after all, one of the many commentators that has argued that the case presents the Supreme Court a binary choice of two bad options – and that what is needed is a much more nuanced policy response. But, unfortunately, that is a task left to the legislators. And that Kerr is wrong on the law.
Kerr’s key error is assuming that because the All Writs Act had been used in analogous situations prior to the passage of the SCA, it should be relied on here. But that argument ignores the changes wrought by the SCA, the differences between how some of those pre-SCA cases operated and the reality today, and the lack of resolution on the underlying key issue in the case.
As Kerr points out, in U.S. v. New York Telephone, the Supreme Court concluded that a court could, via a combination of the Rule 41 warrant authority and the All Writs Act, order a telephone company to assist the FBI in operating a pen register. In that case, the FBI sought to operate the pen register, but needed access to facilities and technical assistance from the telephone company in order to do so. The Supreme Court concluded that the company’s assistance was appropriate and required given that it was needed to implement a pen register that “the District Court was empowered to issue by Rule 41.”
Three things are different here:
First, we now have the SCA, which sets out a detailed statutory scheme authorizing the kind of compelled disclosure warrants at issue in the case. Several parts of the SCA support this conclusion.
• 18 USC 2703(a) explicitly states that the “a governmental entity may require the disclosure by a provider of electronic communication service of the contents of a wire or electronic communication . . .only pursuant to a warrant issued using the procedures described in the Federal Rules of Criminal Procedure. . .” (emphasis added). Kerr focuses on the reference to a warrant in support in support of his theory that the SCA isn’t providing any new authority. But simply specifies the procedures that are required –namely the warrant-like requirements of a finding of probable cause by an independent judge. Meanwhile, first part of the provision provides the explicit authority to compel, specifying that the government “may require” the sought-after disclosure.
• Other statutory provisions make the fact that the SCA provides a new form of authorization explicitly clear, referring to warrants “issued” either “under” or “in accordance” with the statute. Section 2703(e) for example, refers entities that comply with “the terms of a . . warrant. . .under this chapter.” Section 2703 (g) talks about a “a search warrant issued in accordance with this chapter’”. And Section 2706 speaks of a “governmental entity obtaining the contents under section . . . 2703.”
• Moreover, the substance of each of these statutory provisions of the statute, both together and independently, make clear that Congress is creating a new authority to compel. Section 2703(e) provides immunity to providers who comply with a SCA warrant. Section 2703 (g) specifies that, unlike ordinary Rule 41 warrants, the FBI does not need to be present for purposes of execution. And Section 2706 provides for cost reimbursement to the companies. There are new procedures and rules designed for the specific authority provided for by SCA warrants – authorizing the government to serve a warrant on a private party and compel that private party to disclose communications content.
Second, unlike the warrant at issue in the New York Telephone case, the kind of warrant being executed in the Microsoft Ireland case does not require FBI operation at all. Instead, it involves a total delegation to the company to conduct the search and seizure at issue and then turn over the sought-after data to the FBI. This is precisely the kind of novel situation that the SCA anticipates and authorizes, via section 2703 (g), which explicitly specifies that the presence of an officer is not required for the service or execution of the warrant.
Notably, the Supreme Court has defined the All Writs Act as a “residual source of authority to issue writs that are not otherwise covered by statute.” Thus, even if the All Writs Act could have once been relied to authorize the kind of search at issue in the Microsoft Ireland case, the SCA now governs.
Third, Kerr’s argument begs the key underlying question: does Rule 41 authorize the kind of compelled disclosure order at issue in the Microsoft Ireland case? In New York Telephone, the court concluded that the judiciary had the power, via the issuance of a Rule 41 warrant, to order the installation of a pen register. In this case, the Supreme Court still needs to grapple with whether, and to what extent, Rule 41 authorizes the court to order the production of data stored overseas. The All Writs Act only comes into play after the court has answered that underlying question – providing a basis for the judiciary to order a company to assist in the exercise of its established Rule 41 authority. In other words, the All Writs Act doesn’t and can’t avoid the key issue in the case – is this a territorial or extraterritorial exercise of the government’s warrant authority?
And while Kerr’s proposed policy response – that it does if the data belongs to a U.S. person but does not if the data belongs to a non-U.S. person—is the kind of nuanced proposal that I am sympathetic to, it is something that needs to be crafted by the Congress, not the Court. Putting aside the many definitional and implementation problems (Kerr for example does not specify whether foreigners located in the United States qualify as U.S. or non-U.S. persons under his scheme), there simply isn’t doctrinal support for his part-territorial, part-extraterritorial solution. After all, the court has never before said that the Rule 41 authority turns on whether the target is a U.S. citizen or a foreigner. While this may very well be the right solution, it is something that needs to – and absolutely should – be decided by the Congress, not the Supreme Court.