Jeffrey Sterling, James Risen, and Prosecuting Leakers: Lessons from the Sterling Trial

UPDATE 1/26/15 3:30 pm – the jury has found Sterling guilty on all nine counts.

UPDATE 5/11/15 – the judge today sentenced Sterling to 42 months in prison.

The long legal saga involving former CIA officer Jeffrey Sterling is drawing to a close. At this writing his criminal trial has concluded and the jury in Alexandria, Virginia is deliberating.

CIA logo

Sterling was indicted on multiple counts of violating the Espionage Act and related charges. The government alleges he illegally leaked classified information concerning a covert CIA operation to New York Times reporter James Risen. The program, dubbed “Operation Merlin,” involved using a Russian scientist working with the U.S. to deliver flawed nuclear weapon blueprints to Iran in an attempt to derail their weapons program.

When Risen first received the information the Times honored a government request not to run the story due to national security concerns. Risen, however, later revealed the same information in his book, State of War.

Ever since Sterling was indicted in 2010, a central drama in the case had been the government’s efforts to compel Risen to testify about whether Sterling gave him the classified information. Risen refused to comply with the government’s subpoena, claiming a reporter’s privilege to refuse to identify his source.

The trial judge initially upheld Risen’s privilege claim and the government appealed. Relying on the 1972 Supreme Court case of Branzburg v. Hayes, the U.S. Court of Appeals for the Fourth Circuit ruled there is no privilege that allows a reporter to refuse to testify in a criminal case. The U.S. Supreme Court declined to consider Risen’s appeal.

After his appeals were exhausted, Risen continued to say he would go to jail rather than comply with the court order to testify and reveal his source. Shortly before the Sterling case went to trial, the government announced that it would withdraw its subpoena of Risen and would not seek his testimony. The trial proceeded without Risen taking the stand.

So what lessons can be drawn from the Sterling/Risen drama?

DOJ’s Decision to Drop the Risen Subpoena 

The big surprise in the case was the government’s decision to drop the Risen subpoena. Risen was a critical witness, and the case was delayed for nearly four years while the government fought for the right to compel him to testify. Why would you do that if you’re not willing to follow through?

The prosecutors ended up going to trial with one hand tied behind their backs. The best possible witness to the alleged disclosure of classified material – the person to whom it was disclosed – did not have to take the stand. The government was left to put on a circumstantial case relying on evidence of phone calls, e-mails, and details in Risen’s book that allegedly could only have come from Sterling. The prosecutors who fought for years to compel Risen to testify must have felt they had the rug pulled out from under them.

But the truth is the government simply bowed to the inevitable. When Judith Miller of the New York Times was jailed for contempt in 2005 for refusing to testify about her source in the Valerie Plame/CIA leak case, that was in connection with a grand jury investigation. The grand jury could simply be put on hold while waiting to see whether the contempt penalty would convince Miller to change her mind. After 85 days in jail, Miller ultimately did agree to testify in the grand jury after her source, Scooter Libby, released her from her promise of confidentiality.

But the Sterling case involved a trial, not a grand jury investigation. If Risen were called to the stand and refused to testify (as he clearly would have), he could be jailed for contempt but the trial would have to go on. The judge would not, and could not, put the entire trial on hold for weeks or months to see whether Risen came around. And once the case was over, Risen would have to be released; a witness can’t be incarcerated to coerce him to testify when the proceeding where his testimony is required no longer exists.

In the end, therefore, the government would have had to endure all the fallout from jailing a reporter and the prosecution still would not have had the benefit of Risen’s testimony. Despite the government’s legal victories, Risen ultimately held all the cards so long as he was willing to go to jail for a few days until the trial was over. Given that reality, it made sense for the government to fold.

The Difficulty in Prosecuting Leak Cases 

The Sterling case highlights the problems with leak cases from a prosecutor’s perspective. First, they are notoriously difficult to investigate. Assuming the leak was to a member of the press, DOJ Guidelines will not allow the prosecutor even to attempt to speak to the only direct witness – the reporter – unless all other possible avenues have been exhausted (and maybe not even then).   That means a difficult investigation trying to identify and interview every potential source of the leak, scouring computers, e-mails and telephone records, and otherwise trying to establish proof sufficient to sustain a conviction with largely circumstantial evidence.

In the rare case where the prosecutor does seek information from the reporter, that effort will almost undoubtedly be met with fierce resistance. There will be a lengthy delay as the reporter, backed by top First Amendment lawyers, litigates whether he or she has a privilege to refuse to testify. Delay is bad for a prosecution: memories fade, witnesses become unavailable, and the case generally gets more difficult to prove.

Finally, if the case ultimately goes to trial, the government will be hamstrung in presenting its case. Matt Zapotosky had a nice piece in the Washington Post last week about the challenges of trying leak cases. To prove a leak of classified information the government has to expose at least some material that, by definition, was not supposed to be public. The Sterling case involved CIA officers testifying behind a screen to protect their identities and certain lines of questioning that were deemed out of bounds due to the restrictions on classified information. The government has to walk a tightrope between proving its case and still protecting the confidential information that led them to bring the case in the first place.

Sometimes these obstacles are insurmountable. In a high-profile embarrassment in 2011, the government was forced to drop felony charges against alleged leaker Thomas Drake after the judge ruled that Drake had to be allowed to reveal certain classified information in order to defend himself. Rather than reveal the information during the trial, the government allowed Drake to plead to a single misdemeanor.

No wonder leak prosecutions are relatively rare and are considered almost a no-win proposition by federal prosecutors.

The Hazards of Relying on Leakers

In his book, Risen claimed that Operation Merlin had been botched and may actually have helped Iran. (This was also Sterling’s view, which is one reason he was suspected to be Risen’s source.) CIA officials at Sterling’s trial painted a very different picture, testifying that the program had been working and that its disclosure damaged our efforts to thwart Iran’s nuclear program and may have put the lives of American assets overseas at risk.

It’s probably not surprising that government officials would have a different view of the program. Given its classified nature it may be years, if ever, before we know the full truth. But this does highlight the hazards for journalists of working with a leaker of classified information.

People leak information for all kinds of reasons, some of them noble and some not so noble. The government alleged that Sterling was a disgruntled former CIA employee who had been fired, had filed multiple personnel actions, and had an ax to grind. Such an individual has an incentive to leak information that makes the agency look bad and to slant their disclosures in a certain way.

A prosecutor working with an informant faces an analogous situation. A good prosecutor realizes that an informant may shade the truth or tell outright lies in order to make themselves look less culpable or to try to tell the prosecutors what the informant thinks they want to hear. If I have an informant telling me about a drug operation, I’m going to investigate to death everything they tell me and try to corroborate it with other witnesses and evidence. I don’t ever want to be in a position of having to take just the informant’s word for anything – or to ask a jury to do so. A prosecutor in such a case has to be wary of “falling in love” with his informant and accepting what they say uncritically.

Journalists, too, may “fall in love” with a source. A good journalist will recognize this risk, of course, but his options for corroborating the source are more limited. If a journalist receives illegally leaked classified information, other people are not likely to talk to him about it. He can’t make extensive inquiries without potentially exposing his source. Even if other government officials wanted to provide contrary information or a fuller picture, they may be unable to do so without revealing additional classified material.

When it comes to stories based on classified information there is a risk that a journalist, armed with what appears to be a juicy disclosure, will run a story based on incomplete or even inaccurate information. The journalist may be getting only one perspective and is seeing only what their source wants them to see. A good and professional journalist will do whatever possible to avoid this, but the risks are still there. And of course there are bad and careless journalists out there, just as there are bad and careless prosecutors.

I have no idea where the truth lies concerning Operation Merlin and I’m certainly no apologist for the excesses of the CIA during the “war on terror.” But given the choice between believing Sterling’s account (as reflected in Risen’s book) and that of the career CIA people who testified at his trial, I see no particular reason to believe Sterling. It seems perfectly plausible to me that the program was working and that its disclosure harmed national security. Only the most cynical partisan would argue that, simply because it’s the CIA, the operation must have been flawed and the officials at Sterling’s trial must have been perjuring themselves.

The broader cautionary point is that not all leakers are virtuous whistleblowers interested only in the public good, not all leaks are a good thing or are necessarily truthful, and just because a secret government operation is disclosed in the press doesn’t mean that the press got it right or got the whole story.

NY Times

The Reaction of the Press

 As you might expect, there has been a fair amount of ball-spiking by the media going on since the government dropped the Risen subpoena, with the press praising Risen as a hero and condemning the prosecution. The New York Times wrote an editorial last week called “Lessons of the James Risen Case,” criticizing the Obama administration’s supposed record of “aggressively attacking investigative journalism” and praising journalists like Risen who “stand up” to the government.

As I discussed in a previous post, leak cases are about trying to stop leaks and protect national security, not about attacking journalism. Due to the nature of the cases clashes with the press happen from time to time, but that’s a far cry from saying that journalism itself is under assault. Only a handful of leak cases are brought, and most of those do not involve attempts to subpoena a journalist.

The Times laments that the Risen case resulted in an “atrocious legal precedent” in the Fourth Circuit denying a reporter’s privilege.  But the Fourth Circuit’s decision was not some kind of new, groundbreaking result; it was a relatively straightforward application of Branzburg, the controlling Supreme Court case that the Times neglected to mention.

The Times and others also argue that the Risen incident demonstrates the need for a federal shield law for reporters. Privilege advocates argue that a shield law is essential to ensure that aggressive investigative reporting such as Risen’s will continue. I completely agree about the critical importance of investigative journalism, but this argument always baffles me. The leaks to Risen, as well as all of the leaks over the years concerning things such as Abu Ghraib, secret CIA prisons, Watergate, and so on, took place without a federal shield law. If anything, the fact that leaking has gone on at a robust pace in the absence of a shield law demonstrates that a shield law is unnecessary.

In any event, the proposed shield laws that have been kicking around Congress for the past decade all have contained exceptions where national security is concerned. Congress is understandably not willing to pass a law providing that sensitive classified information may be disclosed to reporters with impunity. Even if a federal shield law were in place, therefore, it likely would not have protected Risen in this case.

There are other significant obstacles to passing a shield law, including how to define who is a “journalist” entitled to the law’s protections. In the end, the Risen case is unlikely to have any significant effect on the long-stalled efforts to get a shield law through Congress.

The press outrage over James Risen and some other recent cases has already prompted the Department of Justice to further tighten the guidelines concerning when a prosecutor may seek information from a reporter. I expect clashes such as the one between Risen and DOJ are going to become even more rare in the future, and leak cases will become even more difficult to prosecute. That may be the real legacy of the Obama administration’s so-called “war on the press.”

“You can’t say that on Facebook!” — The Supreme Court Evaluates On-line Threats in Elonis v. United States

Update 6/1/15:  Today the Supreme Court reversed Elonis’ conviction in a 7-2 decision.  I’ll have an analysis of the opinion in next week’s post.

If a man posts violent threats on his own Facebook wall, could he be convicted of a crime even if he didn’t mean it? That’s the question the Supreme Court took up this week in the “Facebook threats” case, Elonis v. United States. It’s a fascinating look at what happens when old legal doctrines bump up against the modern on-line world.

The Facebook Posts at Issue


The defendant Anthony Elonis was convicted for a series of posts he made on his own Facebook wall in October and November 2010, when he was 27 years old. Elonis had been having an emotionally turbulent year: in May his wife left him, taking their two children with her, and in October he lost his job.

Elonis was active on Facebook; he had hundreds of “friends” and posted about a wide variety of topics. After his wife left him, he began posting some compositions of his own. These were often in the form of rap lyrics, and were frequently crude, graphic and violent.

Along with the violent posts, Elonis frequently posted disclaimers, saying his posts were merely “fictitious lyrics,” were for “entertainment purposes only,” or that he was simply exercising his First Amendment rights. He also regularly linked to things such as the Wikipedia entry on freedom of speech and other articles about the First Amendment.

In the fall of 2010, Elonis’s Facebook posts about his wife became increasingly graphic and violent. One post read in part:

There’s one way to love ya but a thousand ways to kill ya

And I’m not gonna rest until your body is a mess,

Soaked in blood and dying from all the little cuts . . .

In November 2010, based on the threatening posts, Elonis’s wife obtained a protection from abuse (“PFA”) order against him. A few days later, Elonis posted an almost word-for-word adaptation of a comedy sketch that he and his wife had watched together, in which comedian Trevor Moore explains that it’s illegal to say you want to kill the President, but not illegal to explain that it’s illegal to say that. The post read in part:

Did you know that it’s illegal for me to say I want to kill my wife?

It’s illegal.

It’s indirect criminal contempt.

It’s one of the only sentences that I’m not allowed to say.

Now it was okay for me to say it right then because I was just telling you that it’s illegal for me to say I want to kill my wife.

I’m not actually saying it. . . .

Elonis followed up this post with a statement that “Art is about pushing limits. I’m willing to go to jail for my constitutional rights. Are you?” He also provided a hyperlink to the original sketch upon which the post was based.

In another post in November 2010, Elonis referred to the PFA order his wife had obtained:

Fold up your PFA and put it in your pocket.

Is it thick enough to stop a bullet?

. . .

I’ve got enough explosives

To take care of the state police and the sheriff’s department . . .

 Elonis’s wife testified at his trial that she took the Facebook threats seriously and that they made her very afraid for herself and her children. She also testified that she had never known Elonis to listen to rap music.

On November 16, Elonis posted the following:

That’s it, I’ve had enough.

I’m checking out and making a name for myself.

Enough elementary schools in a ten mile radius

To initiate the most heinous school shooting ever imagined.

And hell hath no fury like a crazy man in a kindergarten class.

The only question is . . . which one?

Elonis testified that this post was based on a rap song by Eminem, I’m Back, in which the rapper fantasizes about participating in the Columbine school shooting.

The post about the elementary school led to a visit by the FBI, during which Elonis declined to be interviewed.   After the agent left, Elonis posted another item on Facebook he titled “Little Agent Lady” in which he falsely claimed he had been wearing a bomb when the agent came to his door and fantasized about killing her:

Took all the strength I had not to turn the bitch ghost

Pull my knife, flick my wrist, and slit her throat . . .

These and other posts ultimately led to Elonis being convicted on four counts of threats, one each for threatening his wife, the police, a school, and the FBI agent. He was sentenced to 44 months in prison.

The Threats Statute

Elonis was convicted under Title 18, U.S. Code, section 875(c), which provides:

Whoever transmits in interstate or foreign commerce any communication containing any threat to kidnap any person or any threat to injure the person of another, shall be fined under this title or imprisoned not more than five years, or both.

Elonis’s defense was that he never intended to threaten anyone. He claimed his posts were therapeutic and a form of artistic expression, similar to the rap artists he professed to admire. He argued that the proper interpretation of 875(c) requires the government to prove that he subjectively intended to put the targets of his alleged threats in fear, and that to prosecute him for anything less violates the First Amendment.

The lower courts rejected his argument and upheld his convictions, applying the majority rule that the intent of the person making the threats does not matter. The government needs to prove only that the defendant made a statement under circumstances or in a context where a reasonable person would foresee that the statement would be interpreted as a serious expression of an intent to do harm – even if the defendant didn’t really mean it or know it would be interpreted that way. The rationale is that the fear and disruption caused by an apparent threat takes place regardless of the speaker’s personal intent, and Congress is free to punish those who cause that fear.

The Supreme Court Argument – Chief Justice Roberts Channels Eminem

Supreme Court

This issue in this case is not whether threats on Facebook can ever be prosecuted. It’s settled that “true threats” fall into the narrow category of speech that is not protected by the First Amendment, along with obscenity,defamation, and “fighting words” that incite violence. The issue is what does the government have to prove about the defendant’s state of mind in order to establish that the statements at issue were indeed “true threats.” During the Supreme Court argument four different possible standards emerged (in decreasing order of the level of proof required):

1) The defendant subjectively and specifically intended that the statements would place the target of his threats in fear of being harmed. (This is the standard argued for by Elonis, at least initially.)

2) The defendant knew that a reasonable person, looking at the statements, would be placed in fear of being harmed. (The “knowledge” standard.)

3) The defendant knowingly made the statements with a reckless disregard for whether the recipient would be placed in fear of being harmed. (The “recklessness” standard.)

4) The defendant knowingly made the statements, and regardless of what the defendant personally knew or intended about their effect, a reasonable person looking at those statements would think they were a serious expression of an intent to harm another. (The standard adopted by the lower courts and most other courts, and argued for by the government.)

The Supreme Court case essentially boils down to this: which of these standards should be the law?

At the oral arguments on Monday, the Court immediately honed in on this issue. The lawyer for Elonis seemed to get in trouble early, giving varied and conflicting responses on what the correct standard should be. Throughout the litigation Elonis had consistently argued that the government should have to prove he subjectively intended to cause fear (#1 above), but during the argument his lawyer seemed to back away from that standard and suggest it would be enough if the government had to prove only that the defendant knew a reasonable person would consider the statements to be true threats (#2 above). His inconsistent responses finally led Justice Scalia to comment, “You really have me confused at this point.”

At the same time, most of the Justices did not appear to be buying the government’s argument that it doesn’t matter at all what the defendant intended or knew, only that the statements looked threatening to a reasonable person. Justice Kagan observed that the government essentially was arguing for criminal liability based on negligent speech, and “that’s not the kind of standard that we typically use in the First Amendment.”

In what may be a Supreme Court first, Chief Justice Roberts quoted rap lyrics by Eminem about drowning his ex-wife, and asked the government attorney whether that could be prosecuted. The response was no, because in the context of a musical rap performance, no reasonable person would perceive Eminem reciting those lyrics as an actual threat – it’s just a performance.

 The Context is the Key


As the exchange with the Chief Justice about Eminem highlighted, when it comes to threats, context is everything. Violent rap lyrics that no one would perceive as a threat in the context of a stage performance could most definitely be a threat if those same words were whispered menacingly into the ear of another person. Similarly, if Elonis had used the same language from some of his Facebook posts in a phone call or written letter to his wife, there’s little doubt they would be considered threats.

But what exactly is the proper context when talking about Facebook? Even within Facebook itself, you can imagine a number of different scenarios. For example: Elonis posted only on his own wall; was not Facebook friends with his wife, the FBI agent, or others who were the subjects of his posts; and he did not tag them. Writing on your own wall is somewhat akin to a public event. The posts can be seen by all of your friends, commented on, forwarded, and “liked” by a  potentially unlimited number people.  Is this more like a rap concert performance or more like whispering in someone’s ear?

Suppose Elonis had tagged his wife, or wrote on her wall instead of his own, or sent her a private message through Facebook? Would each of these lead to a different result? And what does a “reasonable person” mean when it comes to interpreting what one sees on Facebook – is it the reasonable, Internet-savvy teenager, or the reasonable septuagenarian Supreme Court Justice?

On-line communication suffers from an inability to convey nuance, tone, inflection, facial expression, body language – all things that can be critical to determining the speaker’s true meaning in face-to-face communication. Probably everyone has had the experience of sending an e-mail or posting something that was intended to be sarcastic or funny but was perceived as serious, or vice-versa. Indeed, a whole world of emoticons has sprung up to help us try to convey emotions or attitudes along with the digital written word.

This makes it particularly important to have legal standards that ensure protected speech does not end up being prosecuted. When it comes to Elonis’s posts, as Justice Scalia repeatedly pointed out, it’s hard to argue that there’s a lot of redeeming social value there.  Nevertheless, the heart of the First Amendment is the protection of even speech that many find vile or offensive, whether it be violent rap lyrics, flag burning, or Ku Klux Klan rallies. And the standard the Court adopts will, of course, affect not merely Elonis but all future speakers (and potential defendants).

The Consequences of a Higher Standard of Proof

I expect the Supreme Court will adopt some kind of a middle ground (#2 or #3 above), not requiring subjective intent as Elonis argued but also not accepting the broad rule requested by the government. Justice Breyer seemed to be suggesting during the argument that #2 was effectively already the law: the defendant has to know that he is making statements that are true threats, which means by definition he has to know that a reasonable person would be put in fear by the statements. The Court may conclude that requiring at least this kind of knowledge is a reasonable middle ground.

This doesn’t mean someone like Elonis could never be prosecuted. As with so many criminal appeals, this case is all about the jury instructions: what should the jury have been told it had to find concerning the defendant’s state of mind? If the Court rules that the Elonis jury should have been told it had to find knowledge or intent by Elonis, his own convictions will be reversed, but for future cases the jury instructions will simply be modified.  Elonis himself could even be re-tried, if the government chose to do so, and I wouldn’t be surprised if he were convicted again.

Proving intent or knowledge is not some kind of insurmountable hurdle. Prosecutors do it all the time. As a prosecutor, I wouldn’t hesitate to take a case like Elonis to a jury and argue that the evidence established he knew or intended that his posts would place his wife in fear. And if a defendant tried to fabricate a “rap lyric” defense, the prosecution could present evidence to establish it was merely a ruse – again, disproving an alleged defense is nothing unusual. If the Court rules for Elonis, it’s not going to be some kind of “get out of jail free” card for future stalkers and harassers.

Imposing a higher proof requirement isn’t about condoning Elonis’s reprehensible conduct, it’s simply about strictly interpreting statutes that criminalize speech. A higher standard of proof will make threats prosecutions somewhat more difficult, but that’s not necessarily a bad thing. The government has to tread very lightly when it seeks to turn written words into a federal felony. Given our First Amendment heritage and devotion to free expression, it’s not too much to ask that the government prove some level of intent or knowledge when seeking to send someone to jail solely for what they wrote. I hope the Court agrees.

What standard should the Court adopt?  Leave a comment below.

Sign up for free e-mail updates of future posts by clicking on the “Follow” button on the upper right.