An Analysis of United States v. Comprehensive Drug Testing: On December 18th, the en banc 9th Circuit will hear oral argument on an important case involving how the Fourth Amendment applies to the search and seizure of computers. You can read the revised panel decision here, and download the petition for rehearing here (5.2mb). I've written a lot about these issues and I worked extensively on them back was at the Justice Department in 1998-2001. Given my writing and past work in the area, I wanted to blog some thoughts about the case.

I. The Facts

  The facts of the case are quite complicated, but here at the basics. The government has been investigating steroid use in baseball, and it obtained warrants for computers owned by the third party company, Comprehensive Drug Testing, that ran the steroid tests. The warrant sought the records of 10 specific players, combined with any "[a]ll manuals, pamphlets, booklets, contracts, agreements and any other materials detailing or explaining" the "administration of Major League Baseball's drug testing program."

  During the execution of the warrant, much of the day was spent in negotiation between the government and CDT as to exactly where the information was located and how it could be copied. The government ended up not taking any equipment, but it did copy a directory of files, called the Tracey directory, that contained a lot of records beyond the mere 10 specific players sought in the warrant. There is some dispute among the parties (and the judges in the panel opinion) as to why the government copied the entire directory rather than stay onsite and only copy the records of the 10 players. Although the facts are kind of unclear on this, it looks like the government ended up looking through the Tracey directory for evidence within the scope of the warrant, and then discovered lots of stuff involving other players' steroid use and sought to expand the investigation on that basis.

  In this case, CDT and the Major League Players' Association have challenged the search warrants and the expanded investigation, arguing on behalf of the other players' interests. Rather than waiting for charges to be filed, which would then lead to a motion to suppress, the groups are trying to assert the Fourth Amendment rights of the unindicted players outside the scope of the initial 10 players using the procedural vehicle of a Rule 41 motion for the return of property unlawfully seized. The question is, should the motion be granted and the investigation preemptively stopped in its tracks? Or should the government be allowed to proceed?

II. Introduction to the Two Big Issues

  This case is particularly interesting because it ends up at the intersection of two distinct problems that I used to deal with a lot when I worked at the Justice Department. As I noted in an earlier article on computer warrants, computer warrants are usually executed in two stages: the physical search stage and the electronic search stage. First, the government goes and takes the files away (the physical stage), and second, the governments searches through the files obtained for the evidence sought by the warrant (the electronic stage). This raises two distinct problems in cases like CDT. At the physical stage, the problem is how to minimize the intrusiveness of the on-site search at operating third-party business. At the electronic stage, the problem is how to minimize the intrusiveness of the off-site search.


Oral Argument in United States v. Comprehensive Drug Testing: I blogged a week or so ago about the Ninth Circuit's pending en banc decision in United States v. Comprehensive Drug Testing, a computer search-and-seizure case arising from the investigation into steroid use in baseball. The Court held its oral argument on Thursday, and the audio is available for downloading here.

  The judges on the en banc panel were Kozinski, Kleinfeld, Graber, Wardlaw, W. Fletcher, Paez, Berzon, Callahan, Bea, M. Smith, and Ikuta. I found it a little hard to get a sense of where the court was going because I could only recognize a few of the judges by their voices. A few panelists expressed strong views, but without knowing who they were it was hard to know if those views were likely associated with outcome-determinative votes.

  Anyway, here are a few general thoughts about the argument:

  1) The issue of whether Rule 41 was the right vehicle for this sort of relief received some attention, but my sense was that most of the judges weren't particularly interested in it. If the en banc court ends up reaching the merits and allows this use of Rule 41, the Ninth Circuit may create a novel sort of pre-indictment suppression remedy that may be the most important development of this case. As a practical matter, only rich defendants would be likely to benefit from such a development: Most searches occur before the right to counsel has attached, so pre-indictment challenges would only tend to be brought by suspects wealthy enough to hire sharp lawyers who would know what to do. But this is still an issue very much worth watching.

  2) At argument, the United States tried to circumvent a lot of the difficult questions raised on the merits by arguing that the common denominator of computer searches was one virtual file, rather than specific information revealed to the government. Because the evidence in dispute here was on one file, counsel argued, there were no issues of plain view. I considered this argument in my article on computer searches, and I ended up rejecting it: I concluded that files are just virtual constructs, and it's arbitrary to base a rule on them. Instead, the common denominator of a computer search should be the information exposed. If a government agent sees part of a file and has to scroll down to the bottom of the file to see the rest, that scrolling amounts to an additional Fourth Amendment search. See pages 554-57 of this article for the argument.

  3) Peters, counsel for the Players' Association, offered what I thought was a highly unrealistic sense of how the government can know when they have the needed evidence responsive to a warrant. If I understood him correctly, Peters suggested at argument that when the company came forward with a piece of paper during the execution of the warrant that the company said was the needed test results, the government should have stopped searching and presumed the company's truthfulness and accuracy. Peters reasoned that it looked like in hindsight that the company was being truthful and accurate, and the government should have recognized that. And at the very least, a very minimalist key word search would have done the job. But knowing something ex post is not the same as knowing it ex ante. I think it's a bad idea to have a Fourth Amendment rule that says that you have to rely on that sort of representation or narrow search when executing a warrant, as it would leave the government in the dark about whether there was a larger picture on the computer that casts doubt on what the government agents think they learned.

  A rule requiring that reliance could also create interesting evidentiary hurdles for the prosecution if charges are brought. Imagine you are counsel for a defendant charged with doping, and your client is being charged largely on the basis of evidence that an employee claimed were the relevant test results. The jury won't know that the government relied on the paper to avoid a Fourth Amendment violation. As a defense attorney, you're going to slam the government for relying on the employee with the piece of paper. The real records are on the computer, you would argue, and the government never even checked the computer! How can your client be convicted on the basis of purported evidence when the government never even confirmed that it was the real thing?

  4) Finally, if the Ninth Circuit ends up ruling against the government, the next important issue will become how the court's new standard interacts with the Fourth Amendment's particularity requirement. The warrants here were facially valid, and the challenge is to their execution. But if the court concludes that the warrants were executed improperly, the government will respond by drafting their warrants more broadly so that their execution is more in line with their facial validity. That will put the new doctrinal pressure on the particularity requirement for computer searches, which courts have so far not enforced very strictly. It will be interesting to see if that might change.

Ninth Circuit Hands Down En Banc Decision in United States v. Comprehensive Drug Testing: The new opinion, on how the Fourth Amendment applies to the search and seizure of a computer in the context of an investigation into steroids in baseball, is here. I'm reading it now and will post some thoughts as I go through it.

Ninth Circuit Enacts Miranda-Like Code for Computer Search and Seizure: The Ninth Circuit's new computer search and seizure decision in United States v. Comprehensive Drug Testing is a truly astonishing decision. The majority opinion, by Judge Alex Kozinski, announces a laundry list of brand-new rules, introduced with no citations to any authority, that henceforth the government must follow when executing warrants for digital information. I can't recall having read anything quite like it, although it does bring to mind Miranda v. Arizona.

   Judge Kozinski helpfully sums up the new rules the Ninth Circuit has announced as follows:
When the government wishes to obtain a warrant to examine a computer hard drive or electronic storage medium in searching for certain incriminating files, or when a search for evidence could result in the seizure of a computer, see, e.g., United States v. Giberson, 527 F.3d 882 (9th Cir. 2008), magistrate judges must be vigilant in observing the guidance we have set out throughout our opinion, which can be summed up as follows:
1. Magistrates should insist that the government waive reliance upon the plain view doctrine in digital evidence cases. See p. 11876 supra.
2. Segregation and redaction must be either done by specialized personnel or an independent third party. See pp. 11880-81 supra. If the segregation is to be done by government computer personnel, it must agree in the warrant application that the computer personnel will not disclose to the investigators any information other than that which is the target of the warrant.
3. Warrants and subpoenas must disclose the actual risks of destruction of information as well as prior efforts to seize that information in other judicial fora. See pp. 11877-78, 11886-87 supra.
4. The government's search protocol must be designed to uncover only the information for which it has probable cause, and only that information may be examined by the case agents. See pp. 11878, 11880-81 supra.
5. The government must destroy or, if the recipient may lawfully possess it, return non-responsive data, keeping the issuing magistrate informed about when it has done so and what it has kept. See p. 11881-82 supra.
  I should add that I think these rules are being announced as Fourth Amendment rules, although at first blush that's not entirely clear: The opinion is remarkably light on the sources of its authority.

  Also, I should add that I'm not sure what most of Judge Kozinski's new requirements actually mean. To pick one example, what does it mean to "waive plain view"? Is the idea that the government promises not to rely on the Fourth Amendment's plain view doctrine to admit evidence out of the scope of the warrant? I have never seen anything like that, and I don't know if such a waiver is even enforceable.

  I'll probably blog a lot about this case over the next few days: This is the most free-wheeling, "look ma no hands" legal decision I've read in a long time, so there is a lot to digest.

Do Magistrate Judges Have a Power to Refuse To Sign Warrants Based on Expectations of How a Warrant Would be Executed?: The Ninth Circuit's new computer search and seizure decision is particularly interesting because of the way it empowers magistrate judges. It envisions magistrate judges as activity overseeing the computer search warrant process, and in particular having the power and duty not to sign warrants unless the judge is satisfied that the warrant will be executed in a way that sufficient protects privacy. The opinion goes out of its way to task magistrates with the job of doing what they need to do to protect privacy:
[W]e must rely on the good sense and vigilance of our magistrate judges, who are in the front line of preserving the constitutional freedoms of our citizens while assisting the government in its legitimate efforts to prosecute criminal activity. Nothing we could say would substitute for the sound judgment that judicial officers must exercise in striking this delicate balance.
  Among the new powers that the Ninth Circuit today specifically bestowed on magistrate judges is this one:
The government should, in future warrant applications, forswear reliance on the plain view doctrine or any similar doctrine that would allow it to retain data to which it has gained access only because it was required to segregate seizable from non-seizable data. If the government doesn't consent to such a waiver, the magistrate judge should order that the seizable and non-seizable data be separated by an independent third party under the supervision of the court, or deny the warrant altogether.
(emphasis added). This raises a very important question that the Ninth Circuit doesn't address, perhaps because it never occurred to the judges on the en banc court: Does a magistrate judge have the power to refuse to sign a warrant that is based on probable cause and is constitutionality particular but that the magistrate judge fears would be executed in a way that is unconstitutional or otherwise too invasive?

  The precedents I am aware of suggest that the answer is "no." See Abraham S. Goldstein, The Search Warrant, the Magistrate, and Judicial Review, 62 N.Y.U. L. Rev. 1173, 1196 (1987) ("The few cases on [whether a magistrate judge can refuse to issue a warrant on the ground that the search may be executed unconstitutionally] hold that a judge has a 'ministerial' duty to issue a warrant after 'probable cause' has been established."); In re Worksite Inspection of Quality Products, Inc., 592 F.2d 611, 613 (1st Cir. 1979) (noting the limited role of magistrate judges in issuing search warrants); Ex Parte United States, 287 U.S. 241, 250 (1932) (holding that a magistrate judge did not have discretion to refuse to issue an arrest warrant after the the grand jury returned an indictment, and noting that "the refusal of the trial court to issue a warrant . . . is, in reality and effect, a refusal to permit the case to come to a hearing upon either questions of law or fact, and falls little short of a refusal to permit the enforcement of the law.").

  I think these authorities make a lot of sense. Under the Fourth Amendment, the magistrate's job is to determine if the warrant is based on probable cause and is sufficiently particular. These are the only requirements of the facial validity of a warrant, and so I would think are the only grounds for a magistrate judge denying a warrant. Otherwise you end up with a tricky situation. Imagine a magistrate refuses to sign a warrant unless the government makes promises as to how the warrant will be executed. On one hand, if the government's promises are enforceable, then the magistrate judge gets to determine the Fourth Amendment as she pleases regardless of what the district judges, circuit judges, and Supreme Court Justices say the law actually is. On the other hand, if the promises the government makes are unenforceable, then you end up with the very awkward situation of the government making a promise to a federal judge and then feeling free to break it.

  Neither of these options are very good ones. But in my view neither is necessary, either: I don't think magistrate judges have the power to condition signing warrants on grounds other than probable cause and particularity. The constitutionality of how the warrant is executed is a question for judicial review after the warrant is executed, not negotiations with a magistrate beforehand. At least that's the traditional understanding: It appears to be no longer the law in the Ninth Circuit as of this morning.

How the Ninth Circuit Tried To End Plain View for Computer Searches Without Ending Plain View for Computer Searches: I think the best way to understand today's remarkable Ninth Circuit Fourth Amendment decision in United States v. Comprehensive Drug Testing is that the Ninth Circuit did its best to end the plain view exception for computer searches without formally ending plain view for computer searches. Chief Judge Kozinski's opinion created an elaborate statute-like new regime to make sure the government acts like there is no plain view exception. The decision is a workaround that effectively ends plain view for computer searches by making sure the government will never collect electronic evidence in plain view in the first place. In this post, I want to explain the basic problem and how the Ninth Circuit took some very unusual steps to solve it.

I. Particularity and Plain View

  First, some context. The Fourth Amendment requires that warrants must particularly describe the place to be searched and the items to be seized. This co-called particularity requirement is designed to limit the scope of searches: Instead of conducting broad, free-ranging searches, the police have to only look in one place for only what the warrant says. One key exception to this is the plain view exception. If the police come across other evidence not described in the warrant, they can seize that evidence as well as long as it is "immediately apparent" that the item is evidence of a crime. The Supreme Court has justified this added power on the ground that if the police are lawfully in a place, they see something that is evidence, and it is so clear that the item is incriminating, they could just get a warrant at that stage anyway. Given that, the thinking goes, it makes sense to let them take it at the time rather than wait for another warrant.

  Critics of the plain view exception argue that it gives the police too much power to conduct general searches. According to critics, the police would say they are searching for evidence A when really they want to poke around for evidence B, C and D, which they suspect is there but don't have enough cause to get a warrant. That would be bad, as it would no longer ensure narrow searches. The Supreme Court has responded that this is unlikely for two reasons. First, the particularly requirement makes this unlikely because it requires police have to conduct narrow searches. If the police they are only conducting narrow searches, they shouldn't have an incentive to look for other stuff. Second, the police can only search where the evidence described in the warrant physically could be located. They can't search in narrow spaces for big things, for example. According to the Court, these realities allow the plain view exception despite fears that it will allow general searches.

II. How Computers Change Particularity and Plain View

  Computers change that. Computer evidence can be located anywhere on a very small storage device, and there can be no obvious shortcut for how to find the evidence sought. As a result, a search for evidence means that no place on the computer can be ruled out: A comprehensive search for the evidence in the warrant will bring almost everything into plain view. Further, electronic storage devices contain an incredible amount of information, and with changing technology, tend to store more and more every passing year. Suddenly the particularity requirement doesn't do the work it once did; suddenly the reasons for allowing the plain view exception don't really apply. As I argued in a 2005 article, Searches and Seizures in a Digital World:
For a variety of reasons, computer technologies may allow warrants that are particular on their face to become general warrants in practice. Computers tend to play an ever greater role in our lives as computer technologies advance, as they are likely to record and store increasingly detailed pictures of our daily experience. At the same time, the particularity requirement does less and less work as the storage capacity of computer devices gets greater and greater. Even if the property described in the warrant is a very specific file or type of information, locating that information may require a broad search for technical reasons.
  The question is, what to do about it? In my article, I discussed a few different possibilities and ended up concluding that eventually courts will have to narrow the plain view exception for computers:
the best way to neutralize dragnet searches is to rethink the plain view exception in the context of digital evidence. The dynamics of computer searches upset the basic assumptions underlying the plain view doctrine. More and more evidence comes into plain view, and the particularity requirement no longer functions effectively as a check on dragnet searches. In this new environment, a tightening of the plain view doctrine may be necessary to ensure that computer warrants that are narrow in theory do not become broad in practice.
  Eventually, I argued, the solution will be to abolish plain view for computer searches entirely. It was too early to take such a step, I argued. But "in time, abolishing the plain view exception may best balance the competing needs of privacy and law enforcement in light of developments in computer technology and the digital forensics process."

III. Understanding United States v. Comprehensive Drug Testing

  The Ninth Circuit did not come out and directly abolish plain view for computer searches in today's case. They really couldn't do that, as there would have been no case or controversy: The search hasn't even happened yet, at least as far as I can tell, so there isn't yet any evidence to exclude and any discussion of plain view directly would have just been dicta.

  Rather than wait until a search has occurred, and then announcing such a dramatic shift then, the EZ Rider & Co. did one better: They created a set of prophylactic rules that has the effect of banning plain view through ex ante restrictions. The Ninth Circuit took a truly remarkable step: It ordered the government to behave exactly as it would behave if the plain view exception did not exist. The court wrote out a list of ways that the government would act if there were no plain view exception, and then ordered the government to follow those rules as a condition of getting the warrant in the first place. The plain view exception is gutted by ensuring the government will never be in a position to try to offer the evidence to a court in the first place.


An Interesting Consequence of United States v. Comprehensive Drug Testing: Am I right that the Ninth Circuit's Fourth Amendment decision in United States v. Comprehensive Drug Testing has rendered every computer search warrant that has ever been obtained — and every offsite search — unconstitutional? I've been working in this area for over a decade, and I have never heard of a case that satisfies the Ninth Circuit's new procedural standards.

  I suppose it's possible that the Ninth Circuit will create a new retroactivity jurisprudence to go along with its new Fourth Amendment jurisprudence. If so, the Ninth Circuit might hold that its new rules don't apply to past warrants or past searches that have occurred. But unless or until it does, I would think that every criminal case in the Ninth Circuit with a search warrant involving computers has just been given a new suppression issue: None of the cases will have followed the protocols that the Ninth Circuit just said the Fourth Amendment requires, as no one could have predicted these new protocols.

  Or perhaps the Ninth Circuit will conclude that there is no suppression remedy for past searches that violated the new requirements? This is really new territory, so it will be interesting to see how it plays out. I suspect we'll find out soon, as there are a lot of these cases.

What Happens When Comprehensive Drug Testing Meets the New Rule 41?: One of the remarkable aspects of the new Ninth Circuit computer search and seizure case, United States v. Comprehensive Drug Testing, is that some of its new rules conflict with the new version of Federal Rule of Criminal Procedure 41 that the United States Supreme Court recently adopted and that is set to go into effect in December. What is going to happen when the new Rule goes into effect? Does the Supreme Court's rule trump the Ninth Circuit's case, or does the Ninth Circuit's case trump the Supreme Court's rule? The answer ends up being a little complicated.

  By way of background, Rule 41 is the rule that regulates federal search warrants. A new version of Rule 41 is set to go into effect in December, and several of its provisions are designed specifically to deal with the new dynamics of computer search and seizures. These rule changes have been in the works for a few years, and the United States Supreme Court adopted the new rules in March. (As an aside, permit me to add with a bit of pride that an article of mine originally helped spur the Rules Committee to address these questions; see Page 13 of this report.)

  It turns out that several of the new changes in the Rules are designed to deal with exactly the problems that the Ninth Circuit tried to solve with yesterday's case. But the two approaches are very different: The United States Supreme Court adopted one approach in March, and the Ninth Circuit then announced a conflicting set of rules yesterday.

  For example, consider the question of what the government is supposed to do when agents execute a search warrant for computers, copy the originals to analyze them, and then return the original. Can they keep the copy that they generated? The new version of Rule 41 says that they can. It states: "The officer may retain a copy of the electronically stored information that was seized or copied." But the Ninth Circuit announced the opposite rule yesterday. Judge Kozinski's opinion states: "The government may not retain copies of such returned data, unless it obtains specific judicial authorization to do so."

  Similarly, consider what kind of notice agents must give the issuing judge as to what was seized, something generally known as the "return" on the warrant. Does the notice need to be of what hardware was seized, or of what data was seized? The new Rule 41 states that the notice need only be of the hardware: "In a case involving the seizure of electronic storage media or the seizure or copying of electronically stored information, the inventory may be limited to describing the physical storage media that were seized or copied." Again, though, the Ninth Circuit's opinion yesterday announced the opposite rule: "within a time specified in the warrant, which should be as soon as practicable, the government must provide the issuing officer with a return disclosing precisely what data it has obtained as a consequence of the search, and what data it has returned to the party from whom it was seized."

  So what happens on December 1, when the new Rule 41 goes into effect? Does the new Rule 41 trump the 9th Circuit decision, or does the 9th Circuit decision trump the new rule?

  Formally speaking, the answer would seem to hinge on whether the new decision was announced as part of the federal supervisory powers authority or whether it was intended as a Fourth Amendment decision. Presumably, a rule handed down by the Ninth Circuit as part of its supervisory powers would give way to a contrary federal rule formally adopted by the United States Supreme Court. If so, then those parts of the new decision that conflict with the new federal rules will have a very short shelf life. On the other hand, if the new rules are intended as constitutional rules, then the Ninth Circuit's rule would trump the new federal Rule.

  So is the source of the Ninth Circuit's new decision the federal supervisory power, or is its source the Fourth Amendment? It's kind of hard to tell, as Ninth Circuit announced the rules with no citations to authority or discussion of where they were getting any of it. As best I can tell, the majority opinion does not bother to identify the source of its authority. At times the opinion mentions the Fourth Amendment, but not with particular seriousness: For the most part the court just handed down the rules.

  One complicating factor is that the Supreme Court has said that courts can't used the supervisory power as a sort of supplement to Fourth Amendment protection. Consider United States v. Payner, 447 U.S. 727 (1980), a case on the interaction between the Fourth Amendment and the federal supervisory power. In that case, the Sixth Circuit used the supervisory power to exclude evidence that the Fourth Amendment did not, as a sort of supplement to Fourth Amendment protection. The Supreme Court reversed, rejecting the use of the supervisory power "as a substitute for established Fourth Amendment doctrine."

  If the new Ninth Circuit decision is in fact a supervisory powers case, then it seems clear that the bulk of it is being used as a substitute for established Fourth Amendment doctrine: As I explained yesterday, the major rules in the case are designed to effectively negate the Fourth Amendment's plain view exception in the context of digital evidence cases. Under Payner, then, that would seem to be unlawful use of the court's supervisory powers (as Judge Ikuta's dissent suggests).

  On the other hand, if the new decision is a constitutional decision, then we end up with a very odd juxtaposition: The new case has several pages of new constitutional rules handed down with no citations based on no particular facts that conflict with rules the United States Supreme Court adopted just a few months ago.

  Either way, this is a pretty remarkable situation.