A few federal court opinions have been making a big public splash recently by taking surprising positions on how the Fourth Amendment applies to location surveillance. The latest opinion in the line is Magistrate Judge James Orenstein‘s decision in In The Matter Of An Application Of The United States Of America And Order For An Order Authorizing The Release Of Historical Cell-Site Information, handed down on Friday. The decision holds that historical cell-cite data — records generated by cell phone providers in the ordinary course of business that indicate which cell towers were communicating with a phone, and thus, the rough location of the phone — is protected by the Fourth Amendment and its warrant requirement.
It’s only a decision by a Magistrate Judge, and it is not binding on anyone. But it is an extraordinary opinion, in my view: It’s an extraordinary result, reached in an extraordinary way, and based on an extraordinary number of errors. In this post, I want to explain the decision and then explain why I think it is plainly incorrect as a matter of Fourth Amendment law.
I. Magistrate Judge Orenstein’s Opinion
The basic reasoning of the opinion follows these steps:
(1) The government has applied for a “specific and articulable facts” court order commanding the phone company to disclose historical cell-site information pursuant to 18 U.S.C. 2703(c). Magistrate Judge Orenstein acknowledges that the government’s application satisfies Section 2703(c). But he states that he has the power to rule on whether a court order, when implemented, will be used in a way that violates the Fourth Amendment — and to deny the order if in his view the statute would be used in an unconstitutional way. He thus makes the application for a court order an opportunity to reach a holding about the Fourth Amendment, and implicitly, the constitutionality of this aspect of the Stored Communication Act.
(2) Magistrate Judge Orenstein then rules that based on the mosaic theory of the Fourth Amendment introduced on August 6 by the D.C. Circuit in United States v. Maynard, that cell-site surveillance should be treated as a search just like GPS surveillance:
I find the opinion in Maynard persuasive, both with respect to its demonstration that Knotts is not dispositive on the issue of prolonged location tracking, and with respect to its analysis of the privacy interest at stake when the government uses technological means to accomplish the kind of prolonged, continuous, and detailed surveillance that would otherwise be impossible.
In contrast to the Maynard court, however, Judge Orenstein does not say exactly what the mosaic is that violates the Fourth Amendment. As a result. he does not say that a certain amount of disclosure would be allowed, but that more is not because it would create a mosaic. Instead, he appears to reason that it is merely “the ability” of the technology to amass the mosaic that triggers the Fourth Amendment. As close as he comes to making this point explicit is in this passage:
The Maynard court’s concern with sustained GPS tracking over the course of a month was not its formally continuous nature, but rather the fact that it results in a vast collection of specific data points that, viewed together, convey the “intimate picture” of the subject’s life. It is the ability to amass a collection of such points, and not the ability to trace the route from each one to the next, that carries with it the ability to resolve those points into a comprehensible picture. Thus, the fact that the government now seeks only certain data points, by itself, does not serve to distinguish the investigative technique at issue here from the one considered in Maynard in any meaningful way.
As best I can tell, the idea here is that historical cell site information has “the ability to resolve those points into a comprehensible picture” that would violate a reasonable expectation of privacy. Thus cell site data is protected by the Fourth Amendment, even in only small amounts, and the order is denied.
(3) Magistrate Judge Orenstein next confronts the closely analogous Supreme Court decision in Smith v. Maryland, which held that numbers dialed are not protected. He distinguishes Smith by relying on a passage from the vacated panel decision from the Sixth Circuit in Warshak v. United States that had held that e-mails contents were protected by the Fourth Amendment. Judge Martin’s opinion in Warshak had taken the view that a reasonable expectation of privacy depends on whether information is likely to be known to be revealed to other people. (This is wrong as a matter of Fourth Amendment law, I should point out, but it’s what the vacated panel decision had suggested.) As best I can tell, Magistrate Judge Orenstein uses that reasoning to suggest that there is a difference between numbers dialed voluntarily when you place a call and the cell towers revealing their contact (not in a way known to the user) in a cell phone call. Thus there is a reasonable expectation of privacy in non-content cell-cite records just like there is in the contents of e-mail in the vacated panel decision in Warshak.
(4) Orenstein then concludes that federal statutory privacy law confirms the reasonableness of an expectation of privacy in cell-site information. To be sure, federal law expressly rejects a warrant requirement for that information in the Stored Communications Act, the law at issue in this case. But there was a law passed in 1999 called the Wireless Communication and Public Safety Act of 1999 that authorized the deployment of a nation-wide 9-1-1 emergency service for wireless phone users, called “Enhanced 9-1-1.” That legislation amended the telecom laws to treat location information as protected information in the telecom setting, which Magistrate Judge Orenstein sees as “legislative approval for the idea that a caller should expect her location information to remain private notwithstanding the unavoidable need to share it with a third-party service provider.”
(5) Judge Orenstein next suggests that the Supreme Court’s third party doctrine is not binding law because the Fourth Amendment must “continually develop.” The third-party doctrine is the rule “that the Fourth Amendment does not prohibit the obtaining of information revealed to a third party and conveyed by him to Government authorities, even if the information is revealed on the assumption that it will be used only for a limited purpose and the confidence placed in the third party will not be betrayed.” United States v. Miller, 425 U.S. 435, 443 (1976). If you accept this language from Miller at face value, Magistrate Judge Orenstein says, then the government is correct that the Fourth Amendment does not apply. But that language can’t be right, he concludes, because such a result would be “irreconcilably at odds with the notion that the Fourth Amendment allows our society to continually develop and refine its definition of the privacy claims it wishes to endorse.” Orenstein explains:
If an explicit declaration by the national legislature that disclosing one’s location information to a service provider should not be deemed to serve as consent for further dissemination of such information, it is difficult to see how our society can effectively implement a consensus to recognize as legitimate an expectation of privacy in any form of information that must be shared among private parties to be usable.
(6) Judge Orenstein recognizes the Supreme Court’s recent warning that “[t]he judiciary risks error by elaborating too fully on the Fourth Amendment implications of emerging technology before its role in society has become clear.” City of Ontario, Cal. v. Quon, — U.S. —-, 130 S. Ct. 2619, 2629 (2010). But he proclaims that there is a flip side, which he explains with this rather dramatic passage at the end:
It likewise risks error in being too reticent on that score. In particular, by waiting too long to weigh in on the constitutionality of warrantless access to newly created kinds of information, the judiciary risks the error of transforming from mere assertion to self-fulfilling prophecy the government’s contention that people categorically lack any reasonable expectation of privacy in such information.
As the Supreme Court acknowledged in Quon when it alluded to the progression from Olmstead to Katz, the Fourth Amendment’s concept of an “unreasonable” intrusion into one’s personal affairs, by its very nature, is not stuck in the amber of the year 1791. That concept must
instead evolve along with the myriad ways in which humans contrive to interact with one another. As the threads that connect us are increasingly entrusted into the hands of strangers who promise to make those connections broader, more intimate, more efficient, and more productive, a jurisprudence that mechanically relies on that fact to disclaim the need for meaningful oversight of the government’s investigative techniques unwisely abandons the critical and continuing task of identifying the expectations of privacy our society is prepared to recognize as reasonable.
The Fourth Amendment cannot properly be read to impose on our populace the dilemma of either ceding to the state any meaningful claim to personal privacy or effectively withdrawing from a technologically maturing society.
II. Why Magistrate Judge Orenstein’s Opinion is Clearly Incorrect on the Law
A lot of people have strong feelings about the propriety of government surveillance as a matter of policy. But as a matter of Fourth Amendment law, I think Judge Orenstein’s opinion is pretty clearly incorrect. It’s incorrect in ways that suggest that Judge Orenstein has a basic misunderstanding of the Fourth Amendment. I think it’s also wrong in misunderstanding the role of magistrate judges in Fourth Amendment law. More broadly, the decision strikes me as the kind of opinion judges write when they have settled on a result they know is a stretch and are marshaling all the support they can, however sparse, to get there. In sum, in my view, the opinion is hard to defend as a neutral application of existing law. In this part of the post, I want to explain why.
For my critique, I’ll follow the series of arguments that Judge Orenstein makes and respond to each:
(1) First, I don’t think Judge Orenstein has the legal authority to rule on the constitutionality of 18 U.S.C. 2703 in an ex ante application. As I have argued in this forthcoming article, applications for court orders do not bestow upon magistrate judges the general authority to imagine how the order would be executed and then to deny the government the power to do what the judge expects would be unconstitutional. Rather, the judge must follow the statute and grant the order. The resulting surveillance can be challenged under the Constitution once it occurs, but the magistrate can’t preemptively adjudicate what hasn’t happened yet when the application is made to him for the court order. Unfortunately, Judge Orenstein doesn’t address this issue: He simply assumes he has the power to do what he does.
(2) Judge Orenstein states that the DC Circuit’s August 6th decision in Maynard is persuasive to him, and announces that he will follow it. I think this is problematic for two reasons. First, given that Maynard is a brand new case from another circuit that is at odds with every other Fourth Amendment surveillance opinion I’ve ever read, as I explained here, it seems weird to just announce that you find it persuasive and to follow it as the leading decision. Why not follow circuit precedent instead? Why not follow the other circuits? Is the new DC Circuit mosaic theory right? It seems to me that some argument for that would be helpful. Not necessarily required, to be sure: But given the state of the law, very helpful.
Second, and more importantly, Judge Orenstein’s opinion actually appears to reject the reasoning of Maynard when it becomes inconvenient. Under the reasoning of Maynard, there is a certain amount of surveillance that is allowed until the combined effect is so much that a mosaic is created. If Judge Orenstein is really following Maynard, he should say that he’ll allow the disclosure of some amount of records but not so many records that a mosaic is created — drawing the line when he thinks a mosaic is drawn. Instead, Judge Orenstein says that it is merely “the ability” of the technology to amass the mosaic that matters. But that’s not what Maynard says: In fact, it’s the opposite of what Maynard says. It seems to me that Orenstein uses the reasoning of Maynard to get around Knotts and to say there was a Fourth Amendment violation, but then he discards the reasoning of Maynard when it would point to the result of allowing non-mosaic evidence to be disclosed without a warrant.
(3) Judge Orenstein’s effort to distinguish the Supreme Court’s decision in Smith v. Maryland is very unpersuasive. As I explained here, a cell site signal is closely analogous to numbers dialed in Smith. It’s a signal that the user sends to the phone company that is necessary for the phone company to deliver the user’s calls. It is a necessary part of placing the call, and information that is necessarily transmitted to the phone company for the course of delivery. It’s information that you don’t want the phone company to have, but that you necessarily have to give them to use the technology: It’s just how the technology works. Given the close similarities, it seems to me that Smith is on all fours with the case of cell-site information. It just seems unavoidable that Smith applies here.
To get around Smith, Judge Orenstein tries to rely on the Sixth Circuit’s vacated panel decision in Warshak, which Judge Orenstein suggests, bizarrely, was somehow retained in part in the Warshak en banc decision (!). What a puzzling choice. The Warshak panel decision was overturned, and its reasoning wiped off the books: Just read the positively blistering dissent that the author of the panel decision (Boyce Martin) filed in his dissent to the Warshak en banc opinion if you’re unsure. Following stray language from the vacated panel decision about what makes an expectation of privacy reasonable is a really poor decision, as that stray language was plainly incorrect as I explained here And it’s not like Warshak was even factually relevant: Warshak involved Fourth Amendment protection for the contents of e-mails, whereas this case involves Fourth Amendment protection for non-content cell-site records of cellular phone calls.
(4) Judge Orenstein next argues that federal privacy law is a source of understandings of what expectations are reasonable. But he has to ignore a lot of law and facts to get there. First, he has to ignore the fact that this argument was tried and rejected by the Supreme Court recently in City of Ontario v. Quon. In Quon, the defendant tried to say that the legal protections of the Stored Communications Act established a reasonable expectation of privacy because it was society’s judgment of what privacy was reasonable. The Supreme Court disagreed:
Respondents point to no authority for the proposition that the existence of statutory protection renders a search per se unreasonable under the Fourth Amendment. And the precedents counsel otherwise. See Virginia v. Moore, 553 U. S. 164, 168 (2008) (search incident to an arrest that was illegal under state law was reasonable); California v. Greenwood, 486 U. S. 35, 43 (1988) (rejecting argument that if state law forbade police search of individual’s garbage the search would violate the Fourth Amendment).
Strangely, Orenstein does not mention Quon, Moore, or Greenwood here. Instead, he reasons from first principles that he thinks statutory privacy laws are guides to what makes an expectation of privacy reasonable. But even if you accepted this (false) premise, the reality is that Congress has enacted a law that rejects a warrant requirement for this type of information in the exact setting in which the government is seeking it in this case: 18 U.S.C. 2703(c). Indeed, when legislative proposals to change the law to impose a warrant requirement have been made, Congress has never acted on them because they haven’t had the support to be enacted into law.
Judge Orenstein works his way around this uncomfortable fact by saying that Congress never really had a specific “intent” to regulate cell-site information when it rejected a warrant requirement for such records, and instead turning to the Enhanced 911 Act that amended the telecom laws. But this is pretty silly, in my view. What do the telecom laws have to do with the issue? And why is the fact that the telecom laws treated cell-site information “differently” indicate that there is a societal view that such information should be protected by a warrant requirement when the law Congress did enact clearly rejected a warrant requirement in this setting? It’s all rather strange. Think about it: Judge Orenstein is saying that Congress’s judgment about the privacy of cell-site records in a 1999 law not implicated in this case requires him to say that the privacy law Congress did enact that plainly covers this case is unconstitutional because it doesn’t reflect Congress’s judgment. That makes no sense: If you want to know Congress’s judgment about privacy, why not just follow the privacy law that Congress actually enacted? It is a strange paean to Congress’s judgment to use that judgment to invalidate Congress’s statute.
(5) But wait, it gets even stranger. Judge Orenstein then explains that the third-party doctrine is not binding on him because the Fourth Amendment must “continually develop.” As best I can tell, Judge Orenstein sees Supreme Court decisions as time-bound: They are in force only so long as Judge Orenstein believes that society today is the same as society was in the past. But that’s not how Fourth Amendment law works. When the Supreme Court rules that the Fourth Amendment does or does not protect something, that is a ruling that lower court judges are required to accept. If the Supreme Court decides to overturn its precedents, it can. But that’s not a job for magistrate judges.
I think Judge Orenstein’s critical error here is in imagining that the reasonable expectation of privacy test is descriptive. He seems to think that judges look at “society” and then answer whether that society sees it as reasonable. If that’s the law, each judge gets to create his or her own Fourth Amendment law based on his own view of what “society” is like in any particular time. That would seem to explain Judge Orenstein’s comment that “by waiting too long to weigh in on the constitutionality of warrantless access to newly created kinds of information, the judiciary risks the error of transforming from mere assertion to self-fulfilling prophecy the government’s contention that people categorically lack any reasonable expectation of privacy in such information.” In Orenstein’s vision, as best I can tell, the judiciary needs to make rulings that shape society and determine what privacy rights that society would think are reasonable. If the government conducts lots of warrantless surveillance, eventually the society accepts that and it becomes reasonable. So he needs to step in and say that this surveillance is unconstitutional to ensure that society will develop in a way that will have this surveillance be unconstitutional.
That’s not how the Fourth Amendment works, however. It’s up to the Supreme Court to say what kind of government conduct violates a reasonable expectation of privacy. Lower court judges, including magistrate judges, have to apply the Supreme Court’s understanding of the concept rather than devise their own. More broadly, the reasonable expectation of privacy test is not just descriptive. Rather it is a normative inquiry made by the Supreme Court, in which the Justices set the terms of the framework for specific kinds of cases in their opinions and lower courts are supposed to apply the court’s decisions in a localized fashion. That’s why Judge Orenstein is just wrong to think that he needs to “weigh in” to avoid a self-fulfilling prophecy: Judicial interpretations of the Fourth Amendment may announce what is protected by a reasonable expectation of privacy as a matter of law, but they do not determine it in any factual sense.
(6) Judge Orenstein’s passage at the end — to the effect that “he Fourth Amendment cannot properly be read to impose on our populace the dilemma of either ceding to the state any meaningful claim to personal privacy or effectively withdrawing from a technologically maturing society” — will ring a bell for readers familiar with Fourth Amendment caselaw. That argument, and the passage at the end more broadly, is almost an exact copy of the argument made by Justice Marshall’s dissent in Smith v. Maryland, the pen register case. But that was just a dissent, and the Supreme Court has gone the other way in rejecting that argument. It seems curious for a magistrate judge to make the argument of a dissenting Justice in the most factually analogous Supreme Court decision instead of the argument of the majority.
(7) Finally, I wanted to flag another strange passage in Judge Orenstein’s opinion. In footnote 5, Judge Orenstein has an extensive discussion of why the law here is unclear. I personally think there are two reasons. First, Congress has chosen not to include a statutory suppression remedy for violations of the SCA. Second, defense attorneys haven’t tried to challenge collection of cell site information for some reason (perhaps because of Smith v. Maryland). But Judge Orenstein blames DOJ instead, in part for not appealing magistrate denials of applications for surveillance orders:
[W]hile the government unquestionably knows when its requests are denied, and just as unquestionably has standing to seek review of such denials, it has (with one exception in the Third Circuit) steadfastly chosen to avoid seeking such review.
The result, in this circuit and others, has been an unpredictable legal regime in which an individual’s right to privacy waxes and wanes based on the fortuity of the location in which an investigation is based and of each district court’s system for assigning miscellaneous criminal
duty to its judges.
This is based on a misunderstanding of the law, I think. A significant reason DOJ mostly hasn’t appealed the denials is not because of “standing”, but rather because of the final order doctrine. The law here is very sparse, but the sparse law tends to suggest that denial of ex ante applications are not appealable final orders. See, e.g., United States v. Savides, 658 F. Supp. 1399, 1404 (N.D. Ill. 1987) (“A probable cause determination on an application for a search warrant by a magistrate is not a final order. Simply stated, the government has no right to appeal if it believes the magistrate erred in denying the warrant.“) (emphasis added), aff’d on other grounds sub. nom, United States v. Pace, 898 F.2d 1218, 1230 (7th Cir. 1990).
It’s true that DOJ did appeal such a denial in the Third Circuit despite Savides. But that doesn’t mean it’s lawful to do so, much less “unquestionably” lawful. The lawfulness of the appeal is uncertain: It just happens that no one challenged DOJ’s right to appeal in that case, so it was never briefed. More broadly, the problem here is not with DOJ, but with Congress. When it enacted the SCA, Congress didn’t think through how courts would have a chance to interpret the law and hand down decisions clarifying the law. That’s a problem of statutory design that Congress needs to fix, I think, not the fault of DOJ.