Monday, February 21, 2011

Swiss Announcement that It will Continue to Drag Its Feet, but Marginally Less So (Perhaps), on Double Tax Treaty Administrative Assistance Requests

The Swiss have announced a relaxed, perhaps, interpretation of their Double Tax Agreement ("DTA") Administrative Assistance provisions. The announcement is here. The action was prompted by OECD peer pressure. Basically, the Swiss signal tax cheats (and other criminals who desire, shall we say, discretion) that the Swiss will generally only respond to administrative requests only when the depositor and the financial institution is named in the request.  Often, as in the UBS / U.S. recent spat, the requesting treaty partner will not know the names of the depositors, so the general requirement that the requesting treaty partner name names -- generally -- means that the Swiss can continue to offer secrecy -- generally.

Here is the key paragraph:

Identifying the taxpayer and the holder of the information in administrative assistance requests will generally continue to be by indicating the name and address. The revision should merely underline that administrative assistance procedures should not be doomed to failure because of the DTA provisions being interpreted too formally and that other means of identification should be admissible. Identification via a bank account is also a possibility, although care should be taken in these cases to ensure that a fishing expedition is not involved.
I think the key word is "generally."  The only exception mentioned is that perhaps the Swiss will respond if a particular bank account is identified, which often is not available to the requesting treaty partner.

One of the issues that some have worried about was whether the agreement reached between the U.S. and the Swiss regarding the interpretation of the DTA to permit requests without names but with certain characteristics would become a roadmap for requests for information from Swiss banks other than UBS. The Swiss answer seems to be no, at least generally. I think it fair to say that the interpretation applied for the UBS disclosures to the IRS will not be repeated unless, of course, peer pressure or some other pressure counsels otherwise.  That is the risk for the U.S. tax cheats who want to play the Swiss game.


  1. I think you are misreading “identification via a bank account”
    What Swiss meant was not the account number but certain account(s) parameters making it(them) 1) unique 2) covered by the existing information exchange treaty
    The best example (and probably what Swiss had in mind) was the UBS criminal (not the John Doe summons) case.
    In Feb. 2009 FINMA handed over to the US details on 285 accounts. There was a court hearing on that handover as several account owners appealed FINMA decision.
    The court upheld FINMA’s decision for the accounts meeting 3 parameters (all 3 combined). The court reasoned that any account meeting all those three parameters would violate QIA UBS signed with the IRS and, therefore, any such an account would establish a case of fraud under Swiss law and the US didn’t need to provide any further information to uniquely identify each and every of those accounts.
    That court decision was widely quoted by various tax blogs and attorneys but it didn’t look like any of them read the actual court ruling.
    I think this is what Swiss mean by “identification via a bank account”

  2. Anonymous, thanks for your comments.

    I have apparently not followed the 285 account disclosure as close as you. But my understanding of the subsequent 4,500 disclosures was that they were identified by characteristics. I suppose the concept is same as parameters you discuss, although they may have been different for the 285 than for the subsequent 4,500.

    Are you saying that this new interpretation and application of the Treaty pursuant to OECD peer pressure will permit the treaty partner -- the U.S. in particular -- to do what it did in either of those earlier turnovers where the depositor was not known and indeed the bank accounts were not known, with the U.S. depositors then being identified by parameters or characteristics?

    I did not read the release that way. I hope some of our other readers will clarify that for us.


  3. That's correct that 4500 accounts were identified by parameters. But those parameters were much broader and covered mostly tax evasion (not subject to the treaty at that time) rather than tax fraud (as defined by Swiss law which was subject to the treaty).
    The initial 285 account was disclosed as they all met three parameters combined:
    1) offshore entity with US beneficiary (internal bank Form A stating the account beneficiary for banking purposes)
    2) Form W-8BEN (or UBS equivalent) was filed certifying that the non-US entity is the account beneficiary
    3) At any point from 01/01/2001 the account held US securities
    Only accounts meeting all those three parameters were disclosed as a combination of those three constituted fraud under Swiss law

    The next batch of accounts was merely tax evasion accounts. Only 500 accounts met definition of fraud, even after stretching that definition to the extend not typical under Swiss law (I’m not sure if those 500 accounts included those 285 already disclosed). The rest – 4000 accounts were “tax evasion” account and not subject to the information exchange treaty at that time. So, the problem was not “name and address” for each account holder. The problem was that even with names and addresses those accounts were not subject to disclosure under the treaty at that time (which didn’t cover tax evasion)
    That’s why that US-Swiss-UBS agreement had to be ratified by the Swiss parliament – to effectively amend the existing treaty. While ratifying that amendment, the parliament explicitly stated that it was a one-time amendment and only applicable to UBS case.

  4. So, are you saying that Swiss bank information will be forthcoming on a treaty request that has (1) with the names and some indication of tax fraud or at least evasion OR (2) without the names (let's call it for convenience a John Doe treaty request) but with the parameters 1-3 you list above (which would be an indication of fraud)?

    That is not the way I read the release, but the release could be using code that only the well-informed (you but not me) could understand. The Swiss clearly are at least recognizing the possibility that they will turnover some account information without the depositor's name so long as it is not a fishing expedition. (I thought fishing was the whole purpose any John Doe type procedure.)

    Since I distrust the Swiss to do anything without their fingers crossed, I am suspicious. (I guess that is redundant.)

  5. Swiss-US DTA was modified in 2009 to include tax evasion as a “covered offence”. I can be mistaken but I think its effective day is September 2009. I’m not an expert in tax law, so I’m not sure what effective date means in practice: any account closed a couple of days before the “effective date” is not subject to new treaty? Any account is subject to new treaty if income was generated after “effective day”? Anything else? You certainly have better understanding of those details. But my understanding is that for any account which is not subject to the modified treaty, your “(1) with the names and some indication of tax fraud or at least evasion” is incorrect in respect to tax evasion
    Your “(2) without the names (let's call it for convenience a John Doe treaty request) but with the parameters 1-3 you list above (which would be an indication of fraud)” seems correct to me in all cases
    For the accounts subject to modified treaty (again, I’m not sure what accounts get affected by “effective date”), my understanding is that as long as the requesting party can uniquely identify the account, no name/address is needed even for tax evasion cases. However, I’m not sure how to “uniquely identify the [tax evasion] account” without making it a fishing.
    A request like “all accounts with no W-9 filed” sounds like a fishing as QIA never required W-9 for no US-security trading accounts

  6. Just for clarity, are you saying that the IRS can make a John Doe treaty request (and expect the Swiss to respond in good faith) with parameters "identifying" characteristics of tax evasion (under the new treaty) and Switzerland is now committed to respond for BOTH of the following (a) Swiss banks, such as HSBC with some U.S. presence (thus sort of like UBS) and (b) Swiss banks with no U.S. presence (and thus arguably not subject to the types of pressures exerted on UBS and, as a result, the Swiss government)? And, particularly with respect

    If the latter is the case, then it would seem that the IRS can flush out from every Swiss bank its own data set of U.S. depositors equivalent to the 4,500 flushed out for UBS. I just find it stunning to think the Swiss would have agreed to that without their fingers crossed.

  7. Again, I’m not sure how tax evasion characteristics can be identified without making it a fishing expedition (which is forbidden even under Article 26)
    Requests like “all accounts with no W-9 filed” or “accounts with the non-FATCA banks” won’t work in my opinion. How would you identify a tax evasion account without using two above?

    HSBC is not a Swiss bank, by the way

  8. Anonymous,

    Hopefully, someone will pipe in with good information about how this is expected to work.

    BTW, HSBC does have a private bank in Switzerland - But that was not a good choice because HSBC's primary location is elsewhere. I was just trying to pick off a Swiss bank that also had some US presence, through its corporate structure, that could give the DOJ some nexus to up the ante for the Swiss and the bank and thus perhaps achieve the same result as for UBS. I suspect that, as of now, this new announcement does mean that all other Swiss banks are now fair game for the type of John Doe treaty request processed for the UBS 4,500 as a result of the pressure put on UBS and the Swiss. But coupled with some other nexus for pressure, the John Doe information might be forthcoming. But, again, these are just speculations based on little more than my suspicion that the Swiss are not going to give up their complicit ways so long as they can delay and obfuscate.


Please make sure that your comment is relevant to the blog entry. For those regular commenters on the blog who otherwise do not want to identify by name, readers would find it helpful if you would choose a unique anonymous indentifier other than just Anonymous. This will help readers identify other comments from a trusted source, so to speak.