As I mentioned earlier today, the Supreme Court has handed down the Grubbs case on the legality of anticipatory warrants. The end result: Anticipatory warrants are constitutional, within limits, and Judge Reinhardt’s special restrictions on their use have been eliminated. My overall assessment: The majority opinion does less damage than it might have, although I think Justice Scalia makes a mathematical error that might inadvertently alter the probable cause requirement.
I’ve written a lot on anticipatory warrants over at the Volokh Conspiracy and the NYU Journal of Law & Liberty, and the concern I have expressed is a simple one: The Fourth Amendment requires that probable cause must exist when the warrant is issued, not when the warrant is executed. Depending on how such warrants are limited, anticipatory warrants can be warrants that are issued before probable cause exists. I’ll assume the reader is familiar with these concerns; to get up to speed, see here, here, and here.
I thought Justice Scalia’s opinion did a reasonably good job addressing this concern. Here is the key passage:
It should be noted . . . that where the anticipatory warrant places a condition (other than the mere passage of time) upon its execution, the first of these determinations goes not merely to what will probably be found if the condition is met. (If that were the extent of the probability determination, an anticipatory warrant could be issued for every house in the country, authorizing search and seizure if contraband should be delivered—though for any single location there is no likelihood that contraband will be delivered.) Rather, the probability determination for a conditioned anticipatory warrant looks also to the likelihood that the condition will occur, and thus that a proper object of seizure will be on the described premises. In other words, for a conditioned anticipatory warrant to comply with the Fourth Amendment’s requirement of probable cause, two prerequisites of probability must be satisfied. It must be true not only that if the triggering condition occurs “there is a fair probability that contraband or evidence of a crime will be found in a particular place,” Gates, supra, at 238, but also that there is probable cause to believe the triggering condition will occur. The supporting affidavit must provide the magistrate with sufficient information to evaluate both aspects of the probable-cause determination.
I think Scalia is misstating probabilities a bit, as I’ll explain in a minute. My guess is that he means to say that the likelihood that the triggering condition must occur must be high enough that there is probable cause to believe that evidence will be found at the place to be searched. (That’s what he says at the beginning of the quoted paragraph.) As a result, probable cause is satisfied. If that’s what he has in mind, that’s a fair response to the concerns I have. It’s not a perfect answer, I think, as it seems to leave open the period of time before the triggering condition will occur. Give an officer enough time to wait until the triggering event occurs, and lots of triggering events will occur eventually (even if months or years down the road). Still, I imagine that the 10-day limit on when issued search warrants must be executed still applies, so the probable cause must be that the triggering event will occur in the 10-day window. In that case, an anticipatory warrant can only be issued if there is probable cause to believe that at some point in the 10-day window the evidence will be located there. If that’s right, I don’t think this opinion is inconsistent with the probable cause requirement (sorry, Mike).
Now, I said before that I think Scalia is misstating his probabilities a bit. Here’s the problem. The combination of 1) probable cause to believe that evidence will be present if a triggering event occurs, plus 2) probable cause to believe the triggering event will occur, does not equal 3) probable cause to believe that the evidence will be present. (Incidentally, “probable cause” and “fair probability” are the same under Gates.) If there is a 50% chance that evidence will be present if a package is delivered, and there is a 50% chance the package will be delivered, that doesn’t mean that there must be a 50% chance the evidence will be present.
If I’m not mistaken, the chances that the evidence will be present equals
(the chances the package will be delivered) X (the chances evidence will be present if the package is delivered) PLUS (the chances that the package won’t be delivered) X (the chances evidence will be present if the package isn’t delivered).
So if the likelihood that evidence will be present is 40% with the triggering event and 5% without it, and there is a 40% chance the triggering event will occur, then the likelihood when the warrant is issued that the evidence will be located at the place to be searched at some point in the 10-day window is actually 19%, not 40%.
I have a few additional thoughts about some rather strange dicta in the case, but I’ll probably save them for another post.
If the chance of satisfying the conditions sufficient for the anticipatory warrant to be valid are lower than normal for other warrants, then it is harder to execute anticipatory warrants than it is to execute other warrants. Isn’t that good?
[OK Comments: I think you have it backwards. It is the chances evidence exists that is lower for anticipatory warrants: thus the warrants are easier to execute, rather than harder.]
You are mistaken. Assuming the warrant will only be acted on if the triggering event does occur, then the probability of that occurring is irrelevant – the chance that anything will be found when the search takes place is therefore by definition equal to the chance something will be found if the triggering event occurs, because the search can’t take place without the triggering event.
[OK Comments: I don't think I'm mistaken. I believe you're missing the point that probable cause must exist when the warrant is issued, not when it is executed.]
I don’t understand the relevance of the 5%. If the triggering event does not occur, then there is no warrant. It does not matter what chance there is of evidence being present without the triggering event.
As I read the opinion, the probability to be assessed is the chance of the [triggering event] and also the chance of [evidence being present after the triggering event].
But your overall point is correct, I think. A 40% chance of a package being delivered, and a 40% chance of the package’s containing evidence is not a 40% overall chance. It’s a 16% chance. That cannot be what Justice Scalia meant.
I thought Justice Scalia meant that the magistrate must make two probable cause determinations, each of which must be present for the anticipatory warrant to issue. I think that makes sense, and I think it does justice to the Fourth Amendment.
[OK Comments: Re your 1st paragraph, I think that's wrong: If the triggering event never occurs, there is still a warrant. It is just not an executed warrant. It is still an issued warrant, which must be based on probable cause.]
I think Scalia’s point is that in the circumstances of this case (controlled delivery) there is probable cause to believe that the evidence will be at the place to be searched when the warrant is to be executed, and he emphasizes when not if. I believe he even italicized this point in the opinion. So the standard is still probable cause, and not the probability of probable cause (what you and I have both called “probably probable cause.”)
I believe you’re mistaken because I believe that Scalia rejects the idea that at the time of the issuance of the warrant, there must be probable cause that contraband will be present at the location to be searched. Rather, he is quite clear that at time of the issuance of the warrant, probable cause must exist that contraband will be present at the location to be searched when the warrant is EXECUTED.
In other words, I do not believe your distinction between executed warrants and issued warrants is tenable — Grubbs seems to argue that the Fourth Amendment only cares about executed warrants. And I believe this is sensible — if one agrees with Amar that the purpose of the warrant requirement is to give official sanction to the reasonableness of a search and take the issue away from a jury, then all that should matter is what the facts were at the time the search actually occurred.
Of course, this does raise the question why the second probability showing is required at all. In theory, it shouldn’t be. I believe the Court adopted it merely to avoid the specter of anticipatory warrants issued for every house in the country, though I fail to see why that’s so horrible since, again, as long as the facts that the magistrate considered to constitute probable cause exist at the time of the search, the warrant has served its function.
Regardless of whether i’m right that the second probability showing shouldn’t be required, I don’t think it can be argued that Scalia keys the inquiry on whether probable cause will exist when the warrant is executed. Thus, there’s no problem with his probabilities.
[OK Comments: I think you misunderstand Scalia's opinion. The requirement of the Fourth Amendment is that probable cause must exist at the time the warrant is issued. That is why Scalia writes, "They require the magistrate to determine (1) that it is now probable" that the evidence will be present (emphasis in original). "Now" refers to at the time the magistrate is making the determination, not when the warrant is executed.]
What would be the result if the warrant issued for the package met the test of the trigger ( in the case facts herein it was just that a package was delivered)but evidence of another crime was discovered. Lets say the package contained drugs instead of porn. Would the drugs be suppressed because it was not in the “trigger” or would it become incidental to the search that was warranted?? I still fear this is a blank check that circumvents probable cause.
In a less obvious but equally important part of this opinion, the Court totally destroyed a line of Ninth Circuit precedent in the stroke of a pen and without much discussion.
On Page 9 of the slip. op., the Court explains that neither the Fourth Amendment nor Fed. R. Crim. Proc 41(d) requires the warrant to be served at the outset. The Ninth Circuit squarely held the opposite in U.S. v. Gantt, 194 F.3d 987 (9th Cir. 1999) and U.S. v. McGrew, 122 F.3d 847, 850 (9th Cir.1997). Without so much as an “honorable mention,” the McGrew/Gantt line of cases is done for.
Not that their demise is a good or bad thing, but it seems to be going unnoticed amid the flutter about anticipatory warrants.
An interesting opinion from the Georgia Court of Appeals dated yesterday:[an anticipatory search warrant was issued for for Smith's residence] The warrant was to be served if five “triggers” occurred: that Deloach would be searched to ensure that he had no contraband; that he would be observed making contact with one or both of the Curtises; that he would ask one of the Curtises to purchase cocaine using marked money; that the person or persons going to Smith’s residence would be followed; and that the person or persons would be searched upon leaving Smith’s residence and, if drugs were found, the search warrant would be served. Affirmed. SMITH v. THE STATE, A05A2020, 2006 Ga. App. LEXIS 321 March 21, 2006. In other words, the entirety of a typical drug investigation is encapsulated into the triggering condition of a search warrant. Seems rather remarkable to me!
It seems to me that Graham and carpundit are correctly agreeing with Prof. Kerr’s correct observation that Scalia was a bit off base in thinking that two probable causes chained together necessarily result in a probable cause — though they certainly do in Grubbs’s case, where the probabilities are far more than 40%. Where Graham and carpundit start differing from Prof. Kerr — and where I also beg to do so — is in the detail of whether 16% or 19% is the correct figure to use in his example — which is to say, in whether the question is the probability that the contraband will be present at the site (conditioned only on what was known when the warrant was issued) or the probability that the contraband will actually be found at the site (again, conditioned on what is known at issuance). Neither Graham, carpundit, Kerr, Scalia, nor myself are arguing that courts should pay attention to the probability the contrabound will be found if the trigger occurs and hence the warrant is executed; that would be an innapropriate condiationalization. But including the 3% chance that the contraband is present but will not be found because the warrant will not be executed also seems inappropriate.
You claim that I am mistaken in arguing that Scalia emphasizes the time when the warrant is executed. As evidence, you quote his language stating “They require the magistrate to determine (1) that it is now probable that the evidence will be present.” But you omitted the rest of the sentence: “They require the magistrate to determine (1) that it is now probable that (2) contraband, evidence of a crime, or a fugitive will be on the described premises (3) when the warrant is executed.” As the full quote makes clear, at the time of issuance, the magistrate must determine whether there is probable cause that the evidence will be at the location when the warrant is executed (meaning that the triggering event has occurred), not whether there is probable cause that the evidence will be there at some later date, regardless of whether the triggering event has occurred.
[OK Comments: But on your reading, then the rest of the paragraph is wrong, correct? See your comment above.]
Orin,
Keep in mind that in most anticipatory warrant situations, the triggering conditions will (and should) be specific enough to lead a judge to conclude that there is a high probability overall that contraband will be present.
In other words, in a controlled delivery situation (the classic example of an anticipatory wwarrant), where the authorities know exactly what the defendant has “ordered” and when he expects to receive his “package,” the probability of the triggering condition occurring (ie, successful delivery of a package) is usually quite high. As Scalia noted, it is highly unlikely that the defendant will refuse delivery (unless, of course, he senses a trap – which presents a problem for any type of search warrant). So, even if probable cause is a multiple of P(triggering condition) and P(evidence), in most cases, the facts will show a 80 or 90 percent probability of the trigger happening.
In fact, I think there is a much greater danger with “ordinary” search warrants, in which the authorities are essentially making an educated guess as to whether the contraband will remain in the location to be searched. In that case, the probable cause calculation should be some multiple of Probability(that D possessed contraband at all) times Prob(that contraband will remain in location by the time the search is executed). Indeed, where the contraband is fungible (e.g., drugs), you’re not even betting that the specific kilos of drugs your undercover cop saw in an apartment last Tuesday will still be there when you perform the search. Instead, you are betting that your putative drug dealer will possess new drugs (or evidence of drug dealing biz) in said apartment. In other words, you are betting that your drug dealer will continue to engage in his illegal activity at the apartment. I don’t see how that qualifier is any different from a triggering condition (and indeed, where the government controls the triggering condition, it seems to me that probable cause is more likely for an anticipatory warrant than an “ordinary” warrant).
As a mathematician I think I can resolve the probability question here. The idea needed is that of conditional probability. Scalia is saying that the event that needs to be probable is “the warrant will be executed and evidence will be found”. The probability for this is not simply the product of the two a-priori probabilities, since the events are not independent. Rather (“Bayes’s Law”) the probability is the product of the probability that the triggering condition will occur and the probability that the search will be successful given the occurance of the triggering condition. Moreover, the second probability is usually very high — for example if we are talking about a police sting operation.