The Feedback Loop in Christopher Steele’s Dossier

Last week, at least three media outlets have provided new details about the relationship between former MI6 officer Christopher Steele — the author of the Trump dossier — and the FBI. First WaPo reported that Steele had reached a verbal agreement that the FBI would pay him to continue his investigation of Russia’s involvement with Trump after still unnamed Democrats stopped paying him after the election. CNN then reported that FBI actually had paid Steele for his expenses. Finally, NBC reported Steele backed out of the deal before it was finalized. Chuck Grassley just sent a letter to Jim Comey asking for more information about the proposed arrangement with Steele.

I’m with Grassley on this. According to WaPo and NBC, FBI would only have paid Steele after the election, presumably regardless of the outcome; by that point Steele’s research couldn’t affect the outcome of the investigation. Nevertheless, the possibility that FBI may have used information from a Democratically paid oppo researcher does raise questions of propriety. Add in the discrepancies in these three reports about whether FBI did pay for Steele’s work, and Grassley is right to raise questions.

I’m also interested in what the relationship says about the way in which political necessities may have impacted the content of Steele’s dossier. All three reports attribute the termination of any FBI-Steele relationship, at least in part, to Steele’s frustration with the FBI. WaPo goes on at some length, explaining that Steele got pissed when Jim Comey reopened the Hillary investigation on October 28, and then grew angrier after the NYT reported the FBI had not confirmed any link to Russia.

Ultimately, the FBI did not pay Steele. Communications between the bureau and the former spy were interrupted as Steele’s now-famous dossier became the subject of news stories, congressional inquiries and presidential denials, according to the people familiar with the arrangement, who spoke on the condition of anonymity because they were not authorized to discuss the matter.


In October, anticipating that funding supplied through the original client would dry up, Steele and the FBI reached a spoken understanding: He would continue his work looking at the Kremlin’s ties to Trump and receive compensation for his efforts.

But Steele’s frustration deepened when FBI Director James B. Comey, who had been silent on the Russia inquiry, announced publicly 11 days before the election that the bureau was investigating a newly discovered cache of emails Clinton had exchanged using her private server, according to people familiar with Steele’s thinking.

Those people say Steele’s frustration with the FBI peaked after an Oct. 31 New York Times story that cited law enforcement sources drawing conclusions that he considered premature. The article said that the FBI had not yet found any “conclusive or direct link” between Trump and the Russian government and that the Russian hacking was not intended to help Trump.

WaPo doesn’t lay this out in detail, however. Here’s what happened on those days in October:

October 28: Comey informs eight committee chairs he will reopen the investigation, which promptly (and predictably) leaks.

October 30: Having been officially briefed on the dossier, Harry Reid writes Comey accusing him of a Hatch Act violation for releasing the information on Clinton while withholding what we know to be information in the dossier.

October 31, 6:52PM: David Corn publishes story based on dossier.

October 31, 9:27PM: NYT publishes article describing multiple investigations into Russian interference, stating “no evidence has emerged that would link him or anyone else in his business or political circle directly to Russia’s election operations.”

October 31, 10:52PM: NYT edits article, adding “conclusive or direct” as a caveat in the sentence “Law enforcement officials say that none of the investigations so far have found any conclusive or direct link between Mr. Trump and the Russian government.”

Notably, assuming the times in Newsdiffs (from which I got the NYT timing) are correct, Steele had already gone public before the NYT published its article. That suggests he (like Harry Reid) believed his research should be part of a competing public story. And by going public in what was obviously a Democratically-seeded article, Steele likely made it far more difficult for FBI to continue the relationship.

Already, these new timeline details raise questions about the degree to which Steele’s concerns that the Trump Russian investigation should have more prominence than the email investigation may have influenced his work. Even if Jim Comey did do something colossally stupid by announcing the reopening of the investigation, that shouldn’t affect Steele’s interest in providing the best intelligence to the US, regardless of the public impact, unless he was always motivated primarily by his role as campaign oppo researcher.

The pointless Alfa Bank report that nevertheless seems to reinforce the dodgy Alfa server story

But I also wonder whether it relates to the content. Consider report 112, dated September 14. It pertains to “Kremlin-Alpha Group Cooperation.” It doesn’t have much point in a dossier aiming to hurt Trump. None of his associates nor the Russian DNC hack are mentioned. It does suggest that that Alfa Group had a “bag carrier … to deliver large amounts of illicit cash to” Putin when he was Deputy Mayor of St. Petersburg, though describes the current relationship as “both carrot and stick,” relying in part on kompromat pertaining to Putin’s activities while Deputy Mayor. It makes no allegations of current bribery, though says mutual leverage helps Putin “do his political bidding.”

As I said, there’s no point to have that Alfa Bank passage in a dossier on Trump. But it does serve, in its disclosure, to add a data point (albeit not a very interesting one) to the Alfa Server story that (we now know) FBI was already reviewing but which hadn’t been pitched to the press yet. In Corn’s piece, he mentions the Alfa Bank story but not the report on Putin’s ties to it. It may be in there because someone — perhaps already in possession of the Alfa Bank allegations — asked Steele to lay out more about Alfa’s ties with Putin.

Here’s one reason that’s interesting, though. Even aside from all the other reasons the Alfa story is dodgy, it was deliberately packaged for press consumption. Rather than the at least 19 servers that Trump’s spam email was pinging, it revealed just two: Alfa Bank and Spectrum Health (the latter of which got spun, anachronistically, as a DeVos organization that thus had to be tight with Trump). Which is to say, the Alfa story was dodgy and packaged by yet unknown people.

The discovery of direct collusion during the intelligence review of the Russian hack

More interesting still is what happens in the period that — according to public reporting, anyway — Steele was working for free.

Contrary to what Steele’s anger suggests, there was no real evidence of direct Russian ties to Trump outside of the famous PeeGate incident (and even if that happened, he was not a knowing participant). In the first report, there’s a claim that “the Kremlin has been feeding TRUMP and his team valuable intelligence … including Democratic presidential candidate Hillary Clinton,” but the part of the report that purportedly describes that sharing states that the Kremlin file on Hillary “had not yet been made available abroad, including to TRUMP or his campaign team,” seemingly contradicting the claim. A subsequent report describes a Presidential Administration official discussed the “possible release [of the dossier] to the Republican’s campaign team,” but without any confirmation that occurred (or even that Trump knew about it).

A subsequent report includes a claim of a “well-developed conspiracy of co-operation between [Trump’s team] and the Russian leadership managed through Paul Manafort and Carter Page. It continued to suggest a quid pro quo between the Russian hack and a shift on Ukraine and NATO policies. But in subsequent discussions of Manafort and Page’s corruption, it drops this claim entirely. Even when Michael Cohen enters the narrative, its about managing fallout over Manafort’s Ukrainian corruption.

There are claims that Trump was trying to set up business in Russia, followed by repeated descriptions of Russians not succeeding in getting him to do so.

In other words, in spite of the fact that there were some really damning allegations in the reports, the subsequent reporting didn’t necessarily back the most inflammatory aspects of them.

After the election, there’s just one report, dated December 13. That dates it to after the CIA’s leak fest reporting that Putin hacked the DNC not just to hurt Hillary and the US, but also to elect Trump. It dates to after Obama ordered an IC report on the hack. It dates to after John McCain delivered yet another copy of the dossier to FBI. It slightly precedes a Crowdstrike report (also done for free) bumping its formerly non-public “medium” confidence Russia’s GRU hacked the DNC to “high.”

And after previous reports describing Michael Cohen’s meetings as serving to cover up Manafort’s corruption and Page’s non-consummated Rosneft deal, this one alleges “the operatives involved [in the DNC hack] had been paid by both TRUMP’s team and the Kremlin,” the first such allegation. That is, over a month after the election but less than a month before its leak, the kind of detail backing direct collusion reappeared in this report.

Chuck Grassley’s questions

Which brings me back to Grassley’s letter. In addition to asking about payments, whether the agreement ever went into force, and whether and how Steele’s material served as a basis for FBI reports or even warrants, Grassley asks a question I’ve long wanted to know: Why we got this version of the memo, which is obviously just a partial selection of the complete dossier (rather like the Alfa story).

  1. How did the FBI first obtain Mr. Steele’s Trump investigation memos?  Has the FBI obtained additional memos from this same source that were not published by Buzzfeed?  If so, please provide copies.

We will actually learn a lot about the validity of the dossier if we see what other parts got dealt to the FBI, and if so whether the copy released to the public was cherry picked for the most damning information.

22 replies
  1. seedeevee says:

    “None of his associates nor the Russian DNC hack are mentioned. ”

    Any day now we will actually see non-CIA leak fest evidence of this “hack”, right?

  2. greengiant says:

    I would think team Trump would have investigated Focus GPS by now albeit having the FBI in the loop all summer belays that. The Intercept article you linked in November indicates it was shopped around “since summertime”.
    Regarding the alleged Trump hotel marketing (Cendyne) Server ALFA connection. I am no expert so here is a very blue sky thought. Email must be passe by now for some communications. I think one could communicate via telnet and using file transfer protocol for example. Quite a few people have probably rolled their own unique methods. No common email data bases standing around, no email metadata, no gmail monitoring and recording, no Petraeus like communications by “draft”, and finally, question mark ?no IP provider sending google every link to check for blacklist and to be recorded? I suspect that the first link from a personal device would be to a server via VPN or TOR, perhaps with some dynamic IP added in. The same for when a communication server would be sending to a personal device, rather than leave a recordable DNS lookup it would use a similar pathway. In this analogy the Cendyne server would have been a “TOR like” exit node with hardcoded link to an Alfa bank server. Using domain seems a little risque implementation.
    The experts say there is no proof. If the Cendyne server was being used for marketing one would think there would be more than 19 DNS lookups.
    I wonder if the NSA is hoovering metadata at this low level of internet protocol.

    • maybe ryan says:

      Tend to agree that spam or even just ‘marketing emails’ is an unsatisfying explanation for a server that hits just 19 other servers, but hits 2 of those servers repeatedly. This concern was already stated in Foer’s original article:

      >But now this capacious server handled a strangely small load of traffic, such a small load that it would be hard for a company to justify the expense and trouble it would take to maintain it. “I get more mail in a day than the server handled,” Davis says.

      >That wasn’t the only oddity. When the researchers pinged the server, they received error messages. They concluded that the server was set to accept only incoming communication from a very small handful of IP addresses.

  3. drouse says:

    Propriety was one of first things that went out the window in this election.Hell the agents in the FBI New York office were using allegations from Clinton Cash for investigation. However, it is getting to the point where damn protecting their sources and methods. If they have evidence, lets have it. All this turmoil is killing us at home and abroad.

  4. LeMoyne says:

    Two points:
    1) Honey Pot or Sting: Steele might have gotten pissed and backed out of dealing with Comey after he put his thumb on the scale (publicly reopening Clinton investigation only to close it again just before election). Oppo researcher or not, Steele could take Comey’s partisanship in flouting rules like Hatch Act and non-disclosure of investigations as bad faith of any interest in actually pursuing leads Steele was collecting. Perhaps saw possibility of the deal being a sting.
    2) Who pared down the dossier? The dossier could have been cherry-picked to avoid disclosing the leads that are actually being followed here or abroad. Could have been pared down to avoid compromising active operations, sources or methods … just sayin’ this… seems to me that the long form dossier should be around somewhere if it was floated around DC and NYC for months before release as so many people say.

    That said, I am so glad you went on DN! the other day EW… I had quite forgotten this rather more careful and studied analysis site exists. Thanks for all you folks do here.

  5. maybe ryan says:

    Something else has been nagging at me.

    What happened to BMaz?

    Everything okay at EW-Central HQ?

  6. J2 says:

    You left out so much pertinent information about the reveal of the dossier.

    As noted this dossier  story had been out and about for months; no media would bite. None

    How did IC insure that it does become news?

    1 Comey briefs Obama on it

    2 Clapper then, in the interest of fairness, briefs Trump

    3 This makes it news. Who goes there? Buzzfeed

    4 Once Buzzfeed publishes the dossier

    5 Send in the clowns: CNN

    IC forced this dossier out to smear Trump. But of course there’s  no such thing as the Deep State

  7. qpl23 says:

    EW: I hope you won’t mind if I venture into some fanciful territory as a thought experiment.

    Thinking about it from Steele’s point of view, when it became clear to him that his work could affect the outcome of a US presidential election (i.e. after Trump won the GOP nomination, if not earlier) it’s easy to believe he would at least check in with his ex-colleagues in MI6 as to how his influence on the future of the UK’s most important national ally might sit with them. This is what respectable ex-spies are expected to do, right? I’m sure John Le Carre would write it like that anyway.

    Assuming he apprised MI6 of his commission, this seems to create an undeniable opportunity for mutual benefit. While MI6 have the chance to inject information of their choosing into Steele’s reports, with the right kind of information Steele’s org can appear to have the reach and resources of a state agency. One could be forgiven for thinking the respective parties would be foolish not to consider the advantages of such an arrangement, assuming that UK might have an interest in deniably helping some facts emerge in reporting on Trump.

    And then, going another turn around the wheel, a similar argument might apply to MI6. Would it not be remiss of them not to consult with their most senior and powerful intelligence sharing partners, namely CIA, and inquire – as a matter of simple courtesy – if they had an interest in surfacing any particular details of Trump’s activities relating to Russia…

    Of course we’d never know if *this* kind of feedback loop (given CIA/NSA are now engaged in verifying the dossier, we’re told) could have happened, though it’s kind of fun to speculate.. But after having considered such an extreme hypothesis, it hardly seems that much worse to wonder about the gaps in memo numbers – were there really any intermediate memos? – and about the dates – do we have any provenance that establishes the memo dates in the released version are reliable, so that analysis in relation to contemporary events is worthwhile? Or could they have been stuck on afterwards so that occurrences such as the Rosneft deal appear prefigured in the dossier?

    But that way Trumpism lies, right? Best not go there!

      • qpl23 says:

        Well, shopping *a* dossier .. if it was “for months” it wouldn’t be “the dossier” at least not in the exact form that buzzfeed published it – assuming the December 13 entry is correctly dated, for example. My question (and I admit it’s somewhat fanciful!) is as to whether there’s any record of what shape Steele’s dossier was in at the various stages of its growth while being “shopped around.”

    • emptywheel says:

      Oh, I think those are all fair questions. That’s some of what I was trying to get to in this post.

      A friend tried to get David Corn to compare the dossier, as leaked, with what he got (looked at) in October. He wouldn’t do that.

      • qpl23 says:

        Thanks for that.

        Another reason for the number gaps might be Steele’s company wrote unrelated reports between the Trump stuff, occupying the other numbers, I guess.

  8. Charles says:

    the possibility that FBI may have used information from a Democratically paid oppo researcher does raise questions of propriety.


    I am not sure I buy this. The FBI uses paid informants all the time. And, while, yes, it raises a question of whether those sources are reporting what their paymasters want to hear, I have never seen any evidence the FBI had the slightest bit of sympathy for Hillary.  It was James Comey who dumped the Abedin/Weiner material just in time for the election.


    Law enforcement has been incredibly (and baselessly) hostile to the Clintons.  I find it difficult to believe they would help her win an election.

  9. Rayne says:

    In re: “the possibility that FBI may have used information from a Democratically paid oppo researcher does raise questions of propriety.”

    Did this get knocked down, then?

    Mr Steele was initially hired by FusionGPS, a Washington, DC-based political research firm, to investigate Mr Trump on behalf of unidentified Republicans who wanted to stop Mr Trump’s bid for the GOP nomination. The British Broadcasting Corporation (BBC) reported that Mr Steele was initially hired by Jeb Bush, one of Mr Trump’s 16 opponents in the 2016 Republican primary. It was not immediately possibly to verify the BBC’s report.

    He was kept on assignment by FusionGPS after Mr Trump won the nomination and his information was circulated to Democratic Party figures and members of the media. (source: Irish Times, 12-JAN-2017, emphasis mine)

    There’s so much disinfo/misinfo now that I can’t find a clean source ruling out Jeb Bush’s initial involvement with FusionGPS.

  10. Joanne Leon says:

    When all the dossier news was breaking, there was a comment over at Pat Lang’s site by a Brit who was watching the news of it on the BBC. He said that they reported that Steele was upset by the reporting because the dossier had been modified, that someone had added things to it that weren’t his work. I thought that was interesting, though it’s hearsay. They also didn’t specify what parts he disowned. But it could line up with the issue of Buzzfeed having published an allegedly incomplete dossier.  It’s also possible that, since this dossier was circulating for months, Buzzfeed had an older version of it. And it’s also possible that the IC added those bits to the dossier themselves, after Trump unexpectedly won the election and everyone was in a panic and many people had noted that the Russia allegations were backed by no real evidence.

Comments are closed.