Last year, I blogged extensively about the pending Fifth Circuit case on Fourth Amendment protection for cell-site data. Readers may recall that this is the case in which a Magistrate Judge denied an application for a court order to obtain cell-site data on the ground that he thought the Fourth Amendment would be violated by the order’s execution. DOJ appealed on Fourth Amendment grounds arguing that cell-site data is unprotected by the Fourth Amendment; various amici argued that the Magistrate Judge was correct and that cell-site data is protected; and I wrote an amicus brief raising procedural objections and arguing that the issue was not ripe for adjudication and therefore the merits couldn’t be reached.
Today the Fifth Circuit issued its opinion, which is available here: In Re: Application of the United States of America for Historical Cell Site Data. This is a long post, so here’s the bottom line: The majority opinion reached the merits and gave the government a huge win. But the opinion addressed and resolved a lot of issues, so in this long post I want to run through all the important legal questions answered by the Fifth Circuit’s opinion. Following the order in the opinion, I’ll start with the court’s procedural holdings and then turn to the merits. I’ll then offer my own analysis, both of the impact of the court’s opinion and of the persuasiveness of its holdings.
Also, given the number and complexity of the issues, I’m going to stick to addressing the majority opinion by Judge Clement. Judge Dennis dissented primarily on statutory grounds, but I won’t focus on that dissent here.
I. The Procedural Holdings
Here are the holdings of the majority opinion:
1) The dispute is ripe because it is limited only to a pure question of law, specifically, whether “the § 2703(d) order provision was categorically unconstitutional with respect to an entire class of records – historical cell site information – that is covered under the plain text of § 2703(c).” According to the court, the legal issue in the case is a general one: “whether every instance of one particular factual circumstance – § 2703(d) orders for historical cell site information – is unconstitutional.” Resolving that issue does not require facts, the court holds, making the issue ripe:
If we conclude that such orders are not categorically unconstitutional, specific orders within that category certainly may be unconstitutional because of additional facts involved in the case. But we do not need such facts to determine if orders for historical cell site records are per se unconstitutional.
See also note 3: “For our review, it does not matter how any eventual search would be carried out. Of course, if the Government executed the order in an unconstitutional manner, any evidence it obtained might be subject to suppression. But that is not the issue presented here.”
Further, there is the required hardship to a party if adjudication of the issue is delayed. The government itself is the party that suffers the hardship, as “[i]t has already been denied the use of § 2703(d) orders for historical cell site information by the district court” and “this is the only time that the Government can challenge the denial of its order.”
2) Appellate jurisdiction exists under 28 U.S.C. 1291 because an appeal from a denial of a 2703(d) order is “an independent proceeding, not tied to any current criminal case, and that denying or granting the order finally disposes of
3) The Magistrate Judge was permitted to address constitutional questions under Article III because the District Court exercises plenary review of any constitutional decisions the magistrate judge decides. (See Footnote 4)
Ok, so those were the procedural holdings. Now on to the merits, which I’ll present in slightly different order than they appear in the opinion.
II. The Merits
1) Historical cell-site records are not protected under the Fourth Amendment because they are business records unprotected under the third party doctrine. Notably, the court is assuming that the stored records will only be records associated with actual calls, not records of where the phone was when it was on but not in use. The records of where the phone was when calls were made is unprotected under the Fourth Amendment because it was a business record of the phone company — generated by a communication to which the phone company was a party:
The cell service provider collects and stores historical cell site data for its own business purposes, perhaps to monitor or optimize service on its network or to accurately bill its customers for the segments of its network that they use. The Government does not require service providers to record this information or store it. The providers control what they record and how long these records are retained. The Government has neither “required [n]or persuaded” providers to keep historical cell site records. Jones, 132 S. Ct. at 961 (Alito, J., concurring in the judgment). In the case of such historical cell site information, the Government merely comes in after the fact and asks a provider to turn over records the provider has already created.
Moreover, these are the providers’ own records of transactions to which it is a party. The caller is not conveying location information to anyone other than his service provider. He is sending information so that the provider can perform the service for which he pays it: to connect his call. And the historical cell site information reveals his location information for addressing purposes, not the contents of his calls. The provider uses this data to properly route his call, while the person he is calling does not receive this information.
2) Regular readers may recall that the third-party doctrine refers to information “knowingly conveyed,” raising the issue of whether users know that their cell-phones send out location information. Most Volokh readers know that, and the Fifth Circuit holds (following Smith v. Maryland) that subscribers do generally:
[A] user voluntarily conveys such information when he places a call, even though he does not directly inform his service provider of the location of the nearest cell phone tower. Because a cell phone user makes a choice to get a phone, to select a particular service provider, and to make a call, and because he knows that the call conveys cell site information, the provider retains this information, and the provider will turn it over to the police if they have a court order, he voluntarily conveys his cell site data each time he makes a call.
3) The court notes that changing technology can change Fourth Amendment rights, but that the better place to strike that balance when technology is in flux is Congress, not the courts:
“[I]n circumstances involving dramatic technological change, the best solution to privacy concerns may be legislative. A legislative body is well situated to gauge changing public attitudes, to draw detailed lines, and to balance privacy and public safety in a comprehensive way.” Jones, 132 S. Ct. at 964 (Alito, J., concurring in the judgment).
Congress has crafted such a legislative solution in the SCA. The statute conforms to existing Supreme Court Fourth Amendment precedent. This precedent, as it now stands, does not recognize a situation where a conventional order for a third party’s voluntarily created business records transforms into a Fourth Amendment search or seizure when the records cover more than some
specified time period or shed light on a target’s activities in an area traditionally protected from governmental intrusion. We decline to create a new rule to hold that Congress’s balancing of privacy and safety is unconstitutional.
We understand that cell phone users may reasonably want their location information to remain private, just as they may want their trash, placed curbside in opaque bags, Greenwood, 486 U.S. at 40-41, or the view of their property from 400 feet above the ground, Florida v. Riley, 488 U.S. 445, 451 (1989), to remain so. But the recourse for these desires is in the market or the political process: in demanding that service providers do away with such records (or anonymize them) or in lobbying elected representatives to enact statutory protections.
The court then notes that its opinion is limited to a specific set of facts, historical cell site records already generated when the phone was in use:
We do not address orders requesting data from all phones that use a tower during a particular interval, orders requesting cell site information for the recipient of a call from the cell phone specified in the order, or orders requesting location information for the duration of the calls or when the phone is idle (assuming the data are available for these periods). Nor do we address situations where the Government surreptitiously installs spyware on a target’s phone or otherwise hijacks the phone’s GPS, with or without the service provider’s help.
4) Finally, according to the Fifth Circuit, the Third Circuit was wrong when it held in 2010 that magistrate judges have the discretion to require the government to show probable cause to obtain a 2703(d) order. The language of 2703(d) requires the judge to issue or deny the order based only on the specific and articulable facts standard: “Congress has weighed these considerations and set this balance. The text of the statute shows that Congress does not want magistrate judges second-guessing its calculus.”
Because the Magistrate Judge had no discretion to demand probable cause and the Fourth Amendment was not implicated in collecting historical cell-site records, the Magistrate Judge should have granted the applications and issued the orders.
III. My Analysis
There’s a ton in this opinion, but here’s my preliminary take. I’ll start with the merits issues (the ones that most people care about) and then turn to the procedural issues (the ones that are great for law nerds but not of interest to most readers).
1) This is a big win for the government. The Fifth Circuit’s opinion is the first federal appellate decision to squarely address how the Fourth Amendment applies to cell-site records. And the court’s opinion is a thorough and comprehensive endorsement of DOJ’s position. Plus, the Fifth Circuit rejected the Third Circuit’s view of 2703(d) that gave magistrate judges the discretion to demand probable cause. It’s only one circuit court opinion, but it’s an across-the-board win for DOJ.
For what it’s worth, I also think that the Fifth Circuit was correct in its analysis of all of the merits issues, at least based on the assumed set of facts. I explain why I think historical cell-site location is unprotected here; why the court correctly read 2703(d) to hold that magistrates lack discretion to demand probable cause in my amicus brief; and why the court was correct to take a cautious approach and defer to legislative decisionmaking here.
Meanwhile, we’ll see if other circuits agree. The civil liberties amici in this case have filed a brief in the Fourth Circuit in a cell-site case involving the more traditional motion to suppress directly pressing the mosaic argument, and we’ll see if the Fourth Circuit (or other circuits in similar cases) agree.
2) On the procedural issues, I want to start by expressing my gratitude to the Fifth Circuit for directly addressing the issues I raised in my amicus brief. Although I tend to disagree with how the panel resolved the procedural issues, I was very impressed by how Judge Clement’s opinion squarely took them on.
With that said, I have two responses to the panel’s analysis of the ripeness question. First, it seems to me that the panel made the legal issue ripe by assuming a set of likely facts and then applying the Fourth Amendment to it. As best I can tell, the court did this by saying that the legal issue was only whether the order was “categorically” and “per se” unconstitutional, which the court indicated was different from the question of whether that actual order would be carried out in a constitutional or unconstitutional way.
The court doesn’t define what it means for an order to be “categorically” or “per se” unconstitutional. At first blush, it almost sounds like the court is applying the Salerno standard for facial challenges. But on the whole that seems unlikely. First, I believe that under Salerno the issue is whether the statute could be applied constitutionally. In contrast, the issue here was whether a specific class of orders — defined by the court as “2703(d) orders to obtain historical cell site information for specified cell phones at the points at which the user places and terminates a call” — are constitutional. Further, I take it that everyone agrees that there are specific facts in which the orders could be carried out without violating the Fourth Amendment, such as if no records existed because the phone was never turned on. The disputed issue is whether the Fourth Amendment is violated in the ordinary case, not whether there are any facts that could avoid a Fourth Amendment issue.
I’m not sure of this, but it seems like the court’s reference to the “categorical” question has something else in mind, something more like whether such orders will lead to unconstitutional conduct ‘in the routine case.’ So when the court says, “[i]f we conclude that such orders are not categorically unconstitutional, specific orders within that category certainly may be unconstitutional because of additional facts involved in the case,” that seems to me to suggest that there is some ordinary case (the categorical case) and then “additional facts” that can make some orders a special case. At least that’s my best sense of the opinion: I’m curious if other readers agree.
If I’m right in reading the opinion that way, the court has essentially created a ripe dispute by assuming a set of ordinary facts. This is also suggested by the court’s passage noting that the opinion is limited to the specific issue before it:
Recognizing that technology is changing rapidly, we decide only the narrow issue before us. Section 2703(d) orders to obtain historical cell site information for specified cell phones at the points at which the user places and terminates a call are not categorically unconstitutional. We do not address orders requesting data from all phones that use a tower during a particular interval, orders requesting cell site information for the recipient of a call from the cell phone specified in the order, or orders requesting location information for the duration of the calls or when the phone is idle (assuming the data are available for these periods). Nor do we address situations where the Government surreptitiously installs spyware on a target’s phone or otherwise hijacks the phone’s GPS, with or without the service provider’s help.
It seems to me that by defining the “narrow issue” before the court, the court is effectively announcing its facts. For example, by saying that the court’s analysis does not include a case in which the records were generated when the phone was idle, it is effectively saying that its presumed facts involve records only created when the phone was actually used.
Second, I was unpersuaded by the court’s analysis of the hardship to the parties required to establish a ripe constitutional dispute. The Fifth Circuit says that there is a hardship to DOJ because it has already been denied the order: DOJ has been waiting for its order for years with its investigation on hold. But if that’s the hardship, was the dispute ripe for adjudication when Magistrate Judge Smith ruled on it? At the time of Smith’s ruling, there was no hardship to DOJ: The order hadn’t been denied. Does that mean that the dispute wasn’t ripe for adjudication when the initial application was made and then became ripe only when the application was denied? Or is the idea that he hardship standard is always satisfied as to any legal issue when a party seeks an ex parte order, on the theory that a denial of the ex parte order causes a hardship? I’m not sure I know what to think of that.
3) Finally, on the issue of whether Magistrate Judges have Article III power to enter constitutional rulings when they receive ex parte applications, I don’t think the court needed to reach the issue; that issue was prompted only for purposes of interpreting 2703(d), which the court did without recourse to the need to interpret the statute to avoid a constitutional question. But given that the court did reach it, I would point out a significant limitation to the court’s analysis. The court holds in Footnote 4 that Magistrate Judges can adjudicate constitutional issues in that setting because there is plenary review by the district court. But as far as I can tell, that is true only in the very rare case when the magistrate judge denies the order. In the overwhelming majority of cases, the magistrate judge grants the order. Maybe I’m missing something, but I’m not sure how there could be any review by an Article III judge of an ex parte application that is granted. The government is the only party and it has a signed order; can it seek review after it won? It’s the absence of review in the usual case — or at least, what I think is the absence of review in the usual case — that creates the possible Article III problem. (I don’t have a position on the Article III issue; I just think it’s difficult enough to trigger the canon of constitutional avoidance in the context of interpreting 2703(d).)