Last week, a divided three-judge panel of the Fourth Circuit ruled in United States v. Graham that the government must obtain a warrant to obtain from a phone user’s historical cell-site location information (CSLI) from a cell phone provider if the requested information covers “an extended period of time.” The opinion by Judge Davis, joined by Judge Thacker, is an important milestone in the ongoing debate on government surveillance authority in the technology age as well as development of more traditional Fourth Amendment doctrine. Judge Motz dissented. As Orin Kerr notes, it also established multidimensional circuit splits that are ripening for Supreme Court review. Below are my thoughts about several lines of tension and contention raised by the Fourth Circuit opinion.

As an Institute for Justice Short Circuit email blast characterized it:

Does the fact that your cell-phone provider has data on your whereabouts and other information contained in your phone mean the government may access this information without a warrant? Nope. The Fourth Circuit says the Supremes should weigh in, but the third-party doctrine (which has been called the “Lochner of search and seizure law”) does not apply.

That is a good enough top line characterization. Suffice it to say, though, the 132-page set of opinions is more complicated.

The case arose out of an investigation into a string of armed robberies in the Baltimore area. During the course of the investigation, and pursuant to warrants, the police seized two cellphones that they then were able to associate with the defendants. Backed by court orders under the Stored Communications Act (SCA), the police obtained historical CSLI from Sprint/Nextel for periods of 14 and 221 days. While CSLI is less precise than GPS, it does provide a lot of locational data (29,659 and 28,410 respective data points for the defendants).

There were three essential holdings in Graham. I address each in turn.

1. Police Review of Extended CSLI Data Constituted a Search

First, the court found the two warrantless, but court order-backed, police requests for the defendants’ historical CSLI from Sprint/Nextel constituted searches under the Fourth Amendment. The court held that the government conducts a search under the Fourth Amendment when it obtains and inspects a cell phone user’s historical CSLI for an “extended” period of time. Here, the case involved 14 and 221 days’ worth of the defendants’ phones, and both constituted an extended period. Users have a reasonable expectation of privacy in both movements over an extended period of time and movements in private spaces, and cell phone CSLI implicates both.

Moreover, the court held, over Judge Motz’s vigorous dissent, that the third party doctrine of United States v. Miller (bank records) and Smith v. Maryland (pen register) was inapplicable because “cell phone users do not voluntarily convey their CSLI to their service providers.” Rather, the court focused on the automatic signal acquisition process between cell towers and cell phones during a person’s movement to negate voluntariness. As such, the court was unwilling to suggest the defendants had assumed the risk of disclosure of CSLI to the government under these facts.

In dissent, Judge Motz focused on two main factors to reach the conclusion that third party doctrine foreclosed the defendants’ arguments. First, she argued that the majority relied on cases that involved direct conduct by the government (e.g., United States v. Jones, United States v. Karo, Kyllo v. United States) rather than by means of a third party. Second, she argued that the majority gave short shrift to the distinction between meta-information and content-based review. To her, CSLI looks a lot more like a pen register or email metadata than a wiretap or review of email content. She also argued that there is a robust role for legislative, rather than constitutional, regulation of police conduct.

In finding the CSLI collection constituted a search, the Fourth Circuit parted company with both the Fifth (In re Application) and Eleventh (United States v. (Quartavious) Davis, en banc) Circuits. Add to the mix that Judge Koh in the Northern District of California recently denied SCA applications for CSLI on grounds similar to Graham. Davis recently filed a writ of certiorari (identifying Just Security’s Jameel Jaffer as one of Davis’s ACLU attorneys), so that would likely be the quickest vehicle for Supreme Court review. Kerr offers some appellate posture analysis, here.

The court’s search analysis implicates a couple of longstanding Fourth Amendment debates:

An Endorsement of Mosaic Theory. In traditional Fourth Amendment analysis, the court would analyze each act alleged to be a search in isolation. However, under a mosaic theory, a series of government acts of surveillance may be analyzed as a collective to determine whether it reaches a tipping point that would trigger a reasonable expectation of privacy and thus be deemed a search. The Fourth Circuit adopts that approach in Graham, echoing Justice Sotomayor’s concerns about government data aggregation in her Jones concurrence and the D.C. Circuit’s reasoning below when the case was named United States v. Maynard.

Normative versus Situational “Reasonableness” of an Expectation of Privacy. Ever since Katz, courts have struggled with the subjective and objective components of a “reasonable expectation of privacy” that society will recognize. If you ask many of my Millennial students and family members whether they believe all their emails and social media are being monitored, they will say “probably.” A situational view of reasonableness would suggest that the Fourth Amendment shrinks as the reality of surveillance alters expectations.

However, I am firmly in the camp that believes the Fourth Amendment has normative content grounded in democratic notions of civil society and ordered liberty. To me, reasonableness of privacy expectation under the Fourth Amendment flows from questions like: How large must the private sphere be to ensure that civil society is preserved? How much privacy do we need to create the apartness necessary to create democratic stewardship of the state rather than subservience to it? Under this view, Fourth Amendment protections do not constrict based on real world experience or technological capacity, but rather the democratic privacy function must be preserved in new technological contexts.

Kerr makes related arguments advancing his equilibrium adjustment theory when he notes, “Fourth Amendment protection resembles the work of drivers trying to maintain constant speed over mountainous terrain: judges add extra gas when facing an uphill climb and ease off the pedal on the downslopes.”

The Graham opinion comes down firmly on the side of normative reasonableness: “The third-party doctrine is intended to delimit Fourth Amendment protections where privacy claims are not reasonable – not to diminish Fourth Amendment protections where new technology provides new means of acquiring private information.” It then quotes Kerr to suggest the alternative would lead to a “Fourth Amendment that matters less and less over time.”

2. Absent a Warrant, the Searches Were Unreasonable Under the Fourth Amendment

In Davis, the Eleventh Circuit held that even if police collection of CSLI constituted a search, it would be reasonable and therefore did not require a warrant. The Fourth Circuit rejected that reasoning by pointing to the fact that SCA section 2703(d) orders require the police to demonstrate something akin to reasonable suspicion, a standard much less than probable cause. As such, absent a “well-delineated exception” to the warrant requirement, the Graham court “decline[d] here to create a new exception to a rule so well established in the context of criminal investigations.”

3. Defendants Are Not Entitled to Exclusionary Remedy Due to Police Good Faith

Crucially, the Fourth Circuit also held that the defendants were not entitled to the exclusionary rule notwithstanding the Fourth Amendment violation because “the government acted in good faith reliance on court orders issued under the SCA.” The court recited the governing precedent about how the exclusionary rule’s “sole purpose … is to deter future Fourth Amendment violations.” It then engaged in the required cost-benefit analysis weighing culpable police conduct to the heavy cost of disregarding reliable evidence of guilt. The Court unsurprisingly determined that the government was entitled to the good faith exception, and the defendants were therefore without a trial remedy.

The good faith exception to the exclusionary rule analysis bothers me because I do not believe the government should be able to use information obtained in violation of the Fourth Amendment as a basis to convict a defendant of a crime. Fourth Amendment rights run to the person holding them and the Fourth Amendment remedy is not merely a rule of police regulation. I am cognizant of, and lament, the significant costs to law enforcement attendant to a robust exclusionary remedy. But those costs, to me, are a function of the Bill of Rights and the integrity of criminal proceedings rather than the kind of situational judicial cost-benefit analysis called for by current precedent. This is the position taken in good-faith exception dissents by Justice Stevens in Arizona v. Evans and Justice Ginsburg in Herring v. United States.

With that said, the good faith exception to the exclusionary rule incentivizes pro-civil liberties rulings by judges facing important Fourth Amendment issues. The old adage that “hard facts make bad law” is mitigated by often taking the hard facts, and reversal of conviction, off the table in the case at hand. On the other hand, the exception also marginally dampens many defendants’ incentive to litigate a Fourth Amendment issue.

In Graham, the officers likely had probable cause to obtain a warrant for the CSLI: They had eyewitness identifications of Graham as the robber of two fast food establishments. (These witnesses prompted an improbable sentence in the fast food wars: “At the scene, employees of Burger King and McDonald’s identified Graham as the robber.”). Using warrants, the police had recovered guns, clothing, vehicles, and stolen articles. Further, they had obtained warrants for purposes of associating the two cell phones at issue with the defendants. If I were a judge considering an application for a warrant for targeted CSLI during the time windows of the crimes, I would have granted it as to the first 14-day period requested by police. As a factual matter, police no doubt acted in good faith reliance on the court order granted under the SCA’s reasonable suspicion-like standard. But the Fourth Amendment requires that the police make these showings to demonstrate probable cause before obtaining extended CSLI information.

*          *          *

We are going to see more and more cases in which courts struggle to apply Fourth Amendment principles, and distinguish facts, across technological platforms that interact with our daily personal lives. Just think about the intimate human activity — and information held by third parties — associated with your Xfinity, Netflix, FitBit, OnStar, Garmin, Apple Watch, and grocery rewards card accounts. Moreover, these thorny issues will cut across traditional criminal investigations as well as counterterrorism investigations. Congress has only begun to get its arms around the innovations in collection of geolocational data at the Justice Department and other federal agencies.

It is an important time for legal scholarship on these topics. The court cited to two of Kerr’s articles, as well as Renée McDonald Hutchins’ Tied Up in Knotts? GPS Technology and the Fourth Amendment, and David Gray and Danielle Citron’s The Right to Quantitative Privacy. I also recommend Brian Owsley’s work on related topics, especially because he had to address some of them when he served as a federal magistrate.

According to Graham, “[p]eople cannot be deemed to have volunteered to forfeit expectations of privacy by simply seeking active participation in society through use of their cell phones.” That could prove to be the most potentially significant language of the opinion because it could apply to a lot of these other platforms if untethered from the specific facts of Graham.