Bundeszentralamt Fur Steuern (Being the Federal Central Tax Office of the Federal Republic of Germany) v Heis & Ors

The primary question raised by the two applications addressed by this judgment is whether appeals by the applicants, the German Federal Tax Office (“the GTA”) and by Deutsche Bank AG (“DB”) against the rejection by the Joint Special Administrators (“the Administrators”) of MF Global UK Limited (“MFGUK”) of their respective proofs of debt, each of which is described below (respectively, “the GTA Appeal” and“the DB Appeals”, and together “the Appeals”), should be stayed by this court. The stays are sought in each case in order to allow the underlying claim which forms the subject of the proof to be resolved by the specialist German tax or fiscal courts, which both the applicants (for different reasons) contend are the natural forum for the determination of the claims and the forum in which they can be resolved most efficiently.

The GTA has commenced a process which would, if a stay is granted, lead to the determination of the issues in the German fiscal courts (subject to a potentially preliminary point, I shall later explain as to whether such courts can make a substantive ruling). However, it is common ground that if no stay is granted, the GTA Appeal and the DB Appeals will proceed in this jurisdiction: the German courts would defer to that decision and automatically stay any proceedings in their own courts relating to the dispute between MFGUK and the GTA as to MFGUK’s entitlement to the withholding tax refunds which have already been paid to MFGUK. That is so even though further proceedings raising substantially the same issues have been issued by MFGUK (through its Administrators) in the German Fiscal Court of Cologne, in respect of which no stay is sought and which will thus proceed to adjudication (and any appeal) in Germany.

A secondary question which arises if the stays sought are refused is as to the directions to be given for the determination of the appeals in this court. In particular, it will be necessary to consider whether the appeals should proceed in parallel, or whether the DB Appeals should await the outcome of the GTA Appeal.

The questions to be addressed raise interesting issues as to (a) the interplay between the principles ordinarily applicable in determining whether to stay proceedings in favour of a foreign forum and the ordinary presumption and preference in the context of a liquidation process in this jurisdiction for all matters of proof to be dealt with here; and (b) the ‘rule against double proof’, which has been the subject of considerable debate.

The proofs that have been rejected and are now subject to appeal all concern refunds of withholding tax obtained in relation to a series of transactions to which MFGUK was party prior to its administration. Such refunds were obtained from the GTA pursuant to tax refund applications made by MFGUK which were at the time accepted by the GTA, which issued tax assessments (“Original Tax Assessment Notices”) sanctioning payment of such refunds accordingly.

The GTA maintains that all these transactions have characteristics typical of “Cum/Ex” schemes (as they have become known). These are explained in more detail below: for the present it is sufficient to say that the GTA considers that “Cum/Ex” transactions in the context of such schemes have no legitimate purposes and are illegitimate devices specifically designed in order to obtain withholding tax refunds from the German treasury in excess of the withholding tax that had been paid over to it.

Accordingly, the GTA’s position is that the relevant refund claims in respect of withholding tax submitted on behalf of MFGUK in respect of more than one hundred transactions bearing such characteristics were false and wrongfully made.

The GTA has issued four tax assessment notices amending (in point of form) the Original Tax Assessment Notices under which it had sanctioned and effected the payment of withholding tax which MFGUK had claimed. In her witness statement made on 17 August 2018, on behalf of the GTA (“Voigt 1”), Ms Tatjana Voigt (“Ms Voigt”) states that as an administrative employee in the Cum/Ex Division of the GTA she has handled Cum/Ex cases on a day to day basis since October 2014 and has explained that the effect of these so-called “Amended Tax Assessment Notices”, as a matter of German law, is that MFGUK is obliged to repay the amounts it received pursuant to the Original Tax Assessment Notices, together with interest.

The GTA regards the matter as particularly important, and as having systemic significance, because “Cum/Ex” schemes have been much used in Germany, have caused very large losses to the German treasury, and have become a matter of considerable public and German governmental concern (Footnote: ¹).

In Voigt 1 it is stated that the GTA “regards this as tax evasion” (and thus illegal). In her second witness statement dated 20 December 2018 (“Voigt 2”), Ms Voigt has explained further that the controversy surrounding “Cum/Ex” schemes is such that the Federal Parliament of Germany convened an inquiry committee to investigate them.

Given these concerns, and the nature of the arguments under German law that it relies on, the GTA has at all times maintained that its claim (“the GTA Claim”) should be determined in Germany by the German tax courts. It explains that it felt compelled to submit a proof in MFGUK’s administration in order to preserve its rights in light of the hard bar date of 15 January 2018 provided for by a CVA proposal by which all claims were to be submitted. Had its proof not been submitted before this date, the CVA proposal provided that the GTA Claim would have been extinguished.

In the event, the CVA proposal did not proceed: it never became fully effective because not all the conditions precedent to its operation were satisfied (Heis v FSCS

Nevertheless, of course, once submitted, the GTA’s proof fell to be determined by the Administrators in accordance with the (UK) Investment Bank Special Administration

The Administrators rejected each such Proof in its entirety. That was before the Court of Appeal decision in respect of the CVA proposal, which was handed down on 11 June 2018: until that decision the CVA proposal had been considered (and at first instance held) to have become effective. The terms of the CVA proposal required that any appeal against the Administrators’ decision to reject a proof had to be made within 21 days of receiving that decision. Both the GTA and DB filed appeals at the end of February 2018 accordingly. Those are the appeals of which stays, alternatively directions, are now sought.

There are two applications, one by the GTA in respect of the GTA Proof/Appeal; the other by DB in respect of both its appeals. Different considerations apply to the two sets of stay applications (that is, DB’s application for a stay in respect of its two appeals and the GTA’s application for stay of its appeal); and DB emphasised, and it is right to record, that it disputes any alleged liability on its part to the GTA in respect of the transactions in issue. However, the DB stay application is dependent upon the success of the GTA application. If the GTA stay application fails, then DB accepts that its application in respect of both its appeals must also fail; but if the GTA’s application succeeds, DB contends that its appeals should be stayed on the same terms also, albeit for rather different reasons.

Two further matters are of particular relevance (especially in the context of determining the appropriate forum) and may conveniently be addressed before I turn to explain ‘Cum/Ex schemes’ and the GTA’s concerns about them in more detail.

The first is to note that under German law, it is within the GTA’s power to give a decision on MFGUK’s objection to the Amended Tax Assessment Notices. If and when it did so, it would then be for MFGUK, if it wished to pursue the matter further, to file an appeal against that decision by the GTA with the Fiscal Court of Cologne. The Fiscal Court of Cologne is one of the 18 fiscal courts in Germany which are the courts of first instance for tax matters (together, “German Fiscal Courts”). The Federal Fiscal Court or “Bundesfinanzhof” or “the FFC”, is the final German appellate court in cases of this nature. The German Fiscal Courts and the FFC are specialist tax courts.

However, the GTA has not yet formally rejected MFGUK’s objection. This is because such objection would establish proceedings in Germany, and there is a procedural rule of German law that, in order to prevent parallel proceedings, a German court will automatically defer to the court first seized of a matter. Accordingly, it seems likely that if the GTA were to reject MFGUK’s objection before this Stay Application has been decided by this Court, on any appeal by MFGUK, the Fiscal Court of Cologne might as a matter of comity defer to this Court in order to avoid parallel proceedings. The GTA considers that there are compelling reasons why the Fiscal Court of Cologne is the appropriate forum to determine the GTA Claim and is consequently eager to avoid any procedural issues that might hinder its ability to do so. However, if the Stay Application is granted, the GTA will issue its rejection of MFGUK’s objection so that, in the event that MFGUK appealed, matters would then proceed before the Fiscal Court of Cologne; in that event, the stay having been granted there would be no bar to the Fiscal Court of Cologne proceeding to deal with the matter.

The second matter is this: in addition to the claims which are the subject of the Amended Tax Assessment Notices, MFGUK has submitted two further applications for refunds of withholding tax to the GTA (“the Later MFGUK Refund Claims”) amounting to a further €49,890,693.13. These applications are made on substantially the same grounds as underlie the reclaims made in respect of the GTA Claim, albeit in respect of different time periods of share trading. One of these applications was submitted by DB on MFGUK’s behalf on 25 July 2011; the other was submitted by the Frankfurt office of KPMG on 17 July 2012 (after MFGUK had gone into administration). As explained in Voigt 1, these further refund claims have been rejected by the GTA (on 27 June 2018) “for essentially the same legal reasons” as the GTA issued the Amended Tax Assessment Notices that form the basis of the GTA Proof/Appeal.

On 26 July 2018 MFGUK (through the Administrators) filed an objection with the GTA against the rejection of the Later MFGUK Refund Claims. Now that this objection has been rejected, under German law, the German Fiscal Courts are the only forum in which MFGUK could appeal. If MFGUK does appeal, then subject to any differences in the transactions in question, it is common ground that the German

I turn next to a more detailed description of Cum/Ex trading and the grounds of the GTA’s objections to it, since a fuller understanding is necessary both in the context of deciding whether the Appeals are likely to turn on issues of fact or on matters of German tax law and also what overlap there is likely to be between the Appeals and the Later MFGUK Refund Claims (which is much disputed between the Administrators and the GTA (and DB)).

I take the following description of the background to this matter, the way in which withholding tax is levied in Germany and the nature of Cum-Ex trading schemes which have given rise to concern very largely from the GTA’s skeleton argument, with additional references to the evidence in Voigt 1.

Withholding tax is levied in Germany on, inter alia, dividends and interest payments. Where withholding tax is chargeable, the company paying the dividends or interest must withhold the requisite portion of the payment on account of the tax due. For entities domiciled in Germany, the proportion of the payment withheld is 26.375%, comprising 25% on account of withholding tax and a “solidarity surcharge” of 5.5% on the tax withheld. The withheld tax is transferred to the German treasury to satisfy the recipient’s tax liability in respect of the dividend or interest.

Pursuant to the Double-Taxation Convention between the UK and the Federal Republic of Germany (the “DTC”), where a UK-domiciled entity is the beneficial owner of shares in a German company, then dividends paid on those shares are taxed at a rate of 15%. The UK shareholder is therefore entitled to claim a refund from the GTA of the difference between the tax due (15%) and the sum withheld by the company paying the dividends (26.375%).

Prior to the appointment of the Administrators, MFGUK traded in shares of German companies. Between 2009 and 31 October 2011 MFGUK made a number of claims for withholding tax refunds both on its own account and through DB as its agent. These claims were, to a great extent, made through an automated electronic filing procedure (the “EFP”) (Footnote: ³).

Between 16 July 2009 and 6 December 2012, the GTA issued 17 tax assessment notices (the Original Tax Assessment Notices), pursuant to which €49,275,715.93 was paid to MFGUK in relation to refunds of German withholding tax claimed by MFGUK. Under the relevant provisions of German law, tax assessment notices, including those issued under the EFP, are subject to re-examination by the GTA.

Although the use of Cum/Ex schemes had been extensive more generally for considerably longer, prompting concern and enquiry on the part of the German authorities into such use, the GTA only began investigating the refunds made under the Original Tax Assessment Notices in 2013. As a result of its investigation over the course of some four years, the GTA concluded that the relevant withholding tax refunds were unlawfully claimed by MFGUK pursuant to “Cum/Ex” schemes which it regarded as flawed and unlawful.

Cum/Ex schemes involve transactions in the shares of a German company on, or closely around, the date on which it declares a dividend. The position of the GTA is that such Cum/Ex schemes follow the same general pattern of short sales of shares in a German company which were not part of the seller’s existing portfolio to a short buyer shortly on or before the dividend-entitlement date, and a pre-ordained sequence of steps designed to procure withholding tax refunds from the GTA in excess of the amounts of tax actually withheld and paid to the German treasury.

In broad overview, the GTA’s position is that Cum/Ex schemes are both flawed and unlawful: flawed, because (it contends) the buyer in a short sale from a short seller does not in law have economic ownership of the shares which is a precondition of an entitlement to a refund of withholding tax paid in respect of a dividend declared in respect of them; unlawful, because the schemes are colourable devices.

In brief summary, such Cum/Ex schemes characteristically operate in the following way:

Thus, according to the GTA, such Cum/Ex schemes have features in common which are present in the MFGUK transactions in question and which according to the GTA “confirm the suspicion that inflows from which [MFGUK] allegedly is to have obtained refund claims [which] arise from pre-arranged short sales over the dividend ex-date”. To quote from paragraph 32 of Voigt 1, these schemes are:

“artificial structures having no commercial purpose or value other than the aim of procuring the payment of tax refunds by the German tax authorities in excess of the amount of tax deducted…”

A useful schematic explaining in more detail the steps typical of such Cum/Ex schemes has been prepared by the Administrators: I attach a copy marked ‘A’.

As will already be apparent, the GTA Proof/Appeal is a direct claim by the GTA to recover €52m of withholding tax refunds paid by the GTA to MFGUK. In essence, the claims are based on the assertion that the payments by the GTA pursuant to the Original Tax Assessment Notices were in fact procured and then made without legal justification and should be repaid with interest. The Administrators characterise the claims as “in essence restitutionary claims”. However, that may be, the fact remains that they arise under the Abgabenordnung (the Fiscal Code of Germany) (the “FCG”) under the German Income Tax Act (“ITA”) out of tax assessments levied by the GTA in Germany relating to tax refunds claimed under German tax legislation.

The GTA seeks to establish its claims on three main grounds. The first and second grounds of the GTA Claim turn on the interpretation and application of particular provisions of the ITA (and in particular section 20(1) number 1 sentence 4 ITA), which was introduced in 2007. The GTA’s third ground is a more general case that Cum/Ex schemes offend general anti-avoidance provisions in the FCG.

The first ground which the GTA relies on is based on the interpretation of detailed provisions of the ITA. The GTA contends that on the correct interpretation of the relevant provision (quoting again from Voigt 1): (a) only one refund claim is permissible in respect of the withholding tax paid on a dividend, and (b) only the legal

In respect of this first ground, two main questions of German law and statutory construction appear to arise:

The GTA contends that

The second ground of the GTA Claim appears to be based on a broader proposition, necessitating factual investigation of the particular transaction which is sought to be impugned, that even if the short buyer (B in the example) would otherwise qualify as the “economic owner” of the relevant shares before actual delivery, nevertheless it should not be treated as such in the particular circumstances. That appears to be on the basis that (to quote a decision of the FFC in 2014) such was the complexity, intricacy and nature of the arrangements that “the taxpayer was ultimately completely isolated from the commercial effects of the share acquisition by a plurality of agreements with other parties.” In such a context, the FFC, in the case before it in 2014, considered that the series of transactions was so complex and contrived that what it termed the “overall contractual concept” could never have led to a transfer of the beneficial ownership in the shares there in question to the buyer, which never acquired the authority to deal with or dispose of the shares.

The GTA contends that the same artificiality and circularity are evident in this case, and that it cannot be said that the buyer ever acquired the right to a refund of the withholding tax in question.

The third and most expansive basis for the GTA Claim is that the true purpose (as well as effect) of the transactions impugned was abuse of the German withholding tax legislation, in breach of the general anti-abuse rule in section 42 of the FCG. This is likely to depend on a detailed factual investigation to ascertain whether the true purpose of an impugned transaction characterised as a Cum/Ex scheme, or of any of its component steps or any collateral arrangements, was tax evasion or abuse, as distinct from a proper commercial purpose.

The same arguments apply in respect of the Later MFGUK Refund Claims amounting to a further €49,890,693.13, in which MFGUK is seeking refunds not yet paid in relation to share trades that occurred in 2011 and 2012. The transactions which are the subject of the Later MFGUK Refund Claims do not differ materially from those that are the subject of the GTA Claim, except that whereas in the latter the claims to withholding tax were generated electronically, and then paid out, which has generated a claim by the GTA for refunds, in the former the claims were by a written procedure, and no payment out was made so that the claim to withholding tax remains outstanding, and it is MFGUK which is the claimant.

I should record, lest there be any doubt, that the Administrators do not accept that there is any basis for the GTA Claim, as the rejection of the GTA Proof demonstrates. In paragraphs 30 to 32 of his seventh witness statement (dated 19 October 2018, “Pink 7”) one of the Administrators, Mr Michael Robert Pink (“Mr Pink”), has emphasised that

“as matters stand, and although investigations by the Administrators are continuing, the Administrators have not identified any evidence of deliberate tax evasion or abuse, rather the Administrators’ investigations…suggest that MFGUK was participating in arbitrage transactions, seeking to profit from pricing differences in the market prior to and following the dividend entitlement date.

The Administrators do not, moreover, accept that MFGUK was in all the relevant instance the counterparty to a “Short Seller”. This is a matter to be ascertained and proved.

Other matters to be investigated and proved include the assertions (Footnote: ⁴) of “stock circulation”, collateral arrangements relating to the “dividend level”, and alleged “pricing anomalies”… These matters are relevant both to the broad issue of tax evasion and the second formulation of the GTA’s claim in terms of the effect of the asserted collateral arrangements, which the GTA must first prove.”

As to DB, its involvement (and exposure) relates to its acts as MFGUK’s custodian bank for German securities. In relation to the claim underlying the DB Mirror Proof/Appeal, DB acted as custodian for MFGUK in circumstances where MFGUK was the purchaser of the relevant German shares (i.e. it was ‘B’ in the context of the explanation above). In relation to the claim underlying the €127m Appeal, DB acted as custodian for MFGUK in circumstances where MFGUK was the intermediate seller (‘S’) of the relevant German shares (and a covered “short seller”).

DB’s position is that it is exposed to claims (a) for its role in acting as custodian bank for MFGUK as buyer (‘B’ in the example given previously) in signing requests for refunds of German withholding tax on its behalf; and (b) for its role in acting as custodian bank for MFGUK as an alleged short seller (‘S’ in the example previously given) in Cum/Ex transactions in allegedly failing to deduct withholding tax in respect of dividend compensation payments by or on behalf of MFGUK to the buyer, it being the position of the GTA that under German law (since an amendment in 2007) the obligation to deduct withholding tax in respect of such payments is imposed on the German credit institution (DB in these cases) executing the sale transaction for the seller.

DB’s claims are for an indemnity against both such risks: (a) in the case of its risk as custodian bank for MFGUK as buyer, this is the subject of the DB Mirror

From Voigt 1, it appears that the GTA’s principal argument is based on an analysis approved by the Fiscal Court of Hesse in two decisions regarding ‘Cum/Ex’ schemes dated 10 February 2016 and 10 March 2017, on a decision with the same result by the Fiscal Court of Dusseldorf, and on the approach of the FFC in April 2014 (though the latter related to a particular and complicated contractual structure so that the FFC’s approach may not be of direct application more generally).

In the Fiscal Courts of Hesse’s final decision (dated 10 March 2017, 4 K 977/14, EFG 2017, 656), which was not appealed, it was determined that the withholding tax claims had been improperly made and had been properly refused in the context of over-the-counter (“OTC”) transactions concluded “cum-dividend” and delivered “exdividend”. The Fiscal Court’s analysis was summarised in Voigt 1 as follows:

The Fiscal Court of Dusseldorf similarly held that the short buyer of shares traded as part of “Cum/Ex” transactions in 1990 was not entitled to take account of withholding tax credits against his corporate income tax liability.

The FFC has apparently thus far only considered one case relating to Cum/Ex transactions (decision dated 16 April 2014, I R 2/12; BFH/NV 2014, 1813). As it is described in Voigt 1, in that case a German limited liability company (GmbH) claimed the right to withholding tax against its corporate income tax liability. The decision of the FFC was based not on whether a share purchaser generally acquires beneficial ownership in shares in the case of a short sale: the FFC held that it did not have to decide that point as a general proposition because on the facts of the case before it, the complicated contractual structure (which had been initiated and modelled by a credit institution in London) involved a series of ultimately circular transactions in the course of which the German GmbH never had the authority to dispose of the shares. The FFC determined that the “overall contractual concept” was such that it held could never have led to a transfer of the beneficial ownership in the shares in the German GmbH. Since confined to those special facts, it appears that that decision does not of itself establish any general principle. However, according to Voigt 1, the “FFC also indicated that [its analysis] may well apply to ‘Cum/Ex’ transactions in general based on a review of the specialist press from which the FFC gathered that ‘Cum/Ex’ transactions are often structured similarly.”

In addition to these civil decisions, it appears from Voigt 1 that the Regional Court of Cologne (Criminal Division) has held that ‘Cum/Ex’ transactions involving short sales in which no withholding tax is deducted while a corresponding application for the refund of the withholding tax is filed meet the criteria for tax evasion under German law. I am told that the Cologne public prosecutor is indeed presently carrying out investigations into the Cum/Ex transactions undertaken by MFGUK; and that there is also an ongoing case in Frankfurt in which the chief prosecutor has brought a charge of tax evasion against a number of defendants based on ‘Cum/Ex’ transactions, though the relevant court in Wiesbaden has not yet determined whether the charges are admissible.

More generally, according to Voigt 1, the Federal Ministry of Finance has announced that public prosecutors and tax investigators are currently taking legal action against banks, lawyers and financial companies on 417 cases, though under German tax secrecy laws, apparently, no further details can be provided.

Further, Ms Voigt states that the issue of Cum/Ex trading has become a topic of some political controversy in Germany. A so-called “inquiry committee” of the Federal Parliament (“the Bundestag Inquiry Committee”) has stated that it is convinced such transactions are “deliberately concealed tax evasion…”.

As to the MFGUK transactions themselves, the GTA, apparently with the assistance of the Cologne public prosecutor, has carried out a four-year investigation, leading it to conclude in summary as follows, based on its investigations to date (again taking this from Voigt 1):

The GTA submits that in these circumstances, the German Tax Courts, rather than this Court, should determine the matters which are the subject of its appeal. It argues that a stay in favour of the German Tax Courts may indeed be mandatory; but even if not, that given especially the German centre of gravity of these matters, the German public interest in, and the systemic importance of, the issues raised, and the fact that there has as yet been no definitive determination of the lawfulness of withholding tax claims further to Cum/Ex schemes, it is in the interests of justice and right and fair in all the circumstances that a stay should be granted and the GTA Claim be determined through proceedings before the specialist tax courts in Germany. DB maintains that it is only logical that its appeals should also be stayed because the same substantive German law issues arise for determination, and the same “forum factors” favouring Germany for resolution of the substantive German law issues will apply.

I shall elaborate on the reasons on which the GTA and DB more particularly rely in this regard. Although one strand of the GTA’s argument is that a stay is mandatory, and I address that below, that was rejected by the Administrators and not in the end pressed very hard by the GTA. The focus of the hearing was principally on the issue whether I should in any event grant a stay in the exercise of my discretion. So, I start by considering the relevant principles to be adopted in exercising my discretion whether or not to stay the proceedings in this jurisdiction to enable the matter to be remitted to the German Fiscal Courts.

There was no material dispute as to the existence of this court’s jurisdiction to stay proceedings, nor, if the decision is not mandated but discretionary, as to the principles guiding the exercise of such discretion.

By Rule 225(2) of the Investment Bank Special Administration (England and Wales) Rules 2011, which apply to MFGUK in its special administration, CPR 3.1(2)(f) applies to the proceedings; that rule reflects and confirms the court’s inherent power to stay proceedings. For the avoidance of doubt, it was not suggested that the present case raises any of the issues concerning the interplay between the Recast Judgments Regulation and this jurisdiction, since that Regulation applies neither to insolvency nor to tax matters: Articles 1.1 and 1.2(b).

The court’s general jurisdiction was described as follows by Moore-Bick J (as he then was) in Reichhold Norway ASA v Goldman Sachs International [1999] CLC 486 at 491:

“The court’s power to stay proceedings is part of its inherent jurisdiction which is expressly preserved by s.49(3) of the Supreme Court Act 1981. It is exercised under a wide range of circumstances to achieve a wide variety of ends. Subject only to statutory restrictions, the jurisdiction to stay proceedings is unfettered and depends only on the exercise of the court’s

jurisdiction in the interests of justice.”

That decision was affirmed on appeal but with the caveat that since a claimant with a bona fide claim not tainted with abuse, oppression or any vexatious quality, is entitled to sue in England any defendant over whom the court has jurisdiction, that entitlement should not be subject to any restriction greater than the interests of justice can properly justify, so that in the ordinary course stays would only be granted in “rare and compelling circumstances” (per Lord Bingham CJ at [2000] 1 WLR 173 at 183H and186B-C).

That broad approach to jurisdiction, but cautious approach to its exercise, has been followed subsequently in a number of cases, including Klöckner Holdings GmbH v Klöckner Beteiligungs GmbH [2005] EWHC 1453 (Comm); Prifti v Musini Sociedad Anonima de Seguros y Reaseguros [2006] Lloyd’s Rep. I.R. 221; and Curtis v Lockheed Martin UK Holdings Ltd [2008] 1 CLC 219.

In both the Prifti case and the Curtis case, however, the Court focused on a particular factor in transnational cases which may of itself amount to a rare and compelling circumstance sufficient to justify a stay: that is, the risk of inconsistent judgments in two or more jurisdictions. Thus:

In the context of insolvency proceedings, the Court’s jurisdiction is arguably broader:

Nevertheless, as it seems to me, in a case where the person seeking a stay has sought by proof to participate in a fund which is being administered in this jurisdiction, the burden is on that person to persuade the Court that the interests of justice are likely to be better served by a stay, notwithstanding any additional burden to which the Administrators may be subject in having to litigate abroad, rather than as part of the administration process here (see also Rubin v Eurofinance [2013] 1 AC 236 at [167]). Administrators and liquidators are not litigating in their own interests but as persons charged with the collection and distribution of the insolvency estate in the interest of the creditors as a whole. Broader interests are thus at stake; and, further, in such a context, the principle that claims should be adjudicated as part of the liquidation process, and (in the UK) pursuant to the machinery established by the Insolvency Act 1986 for that purpose, is an important, though not conclusive, one.

Thus, in Enron Metals & Commodity Ltd v HIH Casualty & General Insurance Limited [2005] EWHC 485 (Ch), Pumfrey J said, at [4]:

“[F]airness in this context is fairness in the context of the provisional liquidation or liquidation as a whole, and the ascertainment of what is fair necessarily involves a consideration of the interests of the creditors as a whole and of the capacity of the provisional liquidators or liquidators to deal with the burden of the proposed litigation.”

The process of proof, and the provisions for rejection and appeals, are carefully controlled by the Insolvency Act and Rules. These rules seek to achieve not only fairness to the general body of creditors as a whole, but also, in the public interest, the efficient and expeditious administration of an insolvent estate. Thus, there are rules for future debts which can be discounted to a present value (Rule 14.44), for a set-off of mutual claims (Rule 14.24), for the valuation of security (Rule 14.15), and other detailed rules for the quantification of creditors’ claims that enable claims to be quantified and a distribution made effectively and expeditiously. These rules cannot be contracted out of because the rules are

“an aid to an orderly winding up of the estate and this benefits the public interest”.

(See McPherson’s Law of Company Liquidation, 4^th Ed, at 12-037.)

These Insolvency Rules are an important part of the statutory scheme. That is a primary reason why the Court has been reluctant to permit any other process for the adjudication of claims than those set out in the rules regulating the lodging, assessment, valuation, rejection and appeals from rejection, which also stipulate strict time limits for the process.

An example of the reluctance of the Court, in controlling the insolvency process, to permit departure from these rules by adoption of some other process of determining claims against an insolvent estate is provided by a recent decision of JusticeSegal in the Grand Court of the Cayman Islands in Re Wimbledon Fund (in official liquidation), 9 July 2018. Segal J refused to permit a creditor who had brought proceedings in New York to continue there in place of the proof of debt filed by the creditor in the official liquidation. He refused the creditor permission, explaining, at [8]-[9], that he considered the Cayman Islands law-governed proof process likely to be productive of a cheaper and quicker resolution of the dispute than litigation in New York, even though an important issue in that case was the impact of a recent New York court decision on the relevant agreement.

Segal J also held it to be significant in that case that the creditor had lodged an appeal in Cayman against the rejection of its proof. Although there was a discretion to direct that certain issues be decided in a foreign court, he said, at [9(j)]:

“[O]nce the creditor has lodged an appeal within the liquidation proceeding the Court will require a clear demonstration that proceedings in this Court could not justly dispose of the claim or that there would be substantial costs and timing benefits to be derived from the foreign proceedings (and that the Court’s control of the appeal would not be compromised).”

That is so, in my view, even where (as here) the decision to participate by way of proof may have been the product of otherwise being excluded by a time bar imposed by the Administrators. As Lord Sumption JSC stated in Stichting Shell Pensioenfonds v Krys [2015] 2 WLR 289, at [28]-[31]:

“[B]y submitting a proof the creditor obtains an immediate benefit consisting in the right to have his claim considered by the liquidator and ultimately by the court according to its merits and satisfied according to the rules of distribution if it is admitted.”

The principal focus of GTA’s submissions was that in the present case it is in the interests of justice and right and fair in all the circumstances that a stay should be granted and that its claim to be entitled pursuant to the Amended Tax Assessment Notices to recover from MFGUK the amounts of withholding tax it received, together with interest, should be determined through proceedings before the specialist tax courts in Germany, where the Later MFGUK Refund Claims are to be adjudicated in any event.

However, and as indicated above, the GTA also floated in Voigt 1 (though it was barely touched on its skeleton argument) the contention that “as a matter of German law, the German Fiscal Courts have exclusive jurisdiction to determine the validity of German tax claims” and that this mandated a stay as well as providing a further reason for the exercise of discretion in its favour.

By “floated” I mean that my impression was that the GTA took the jurisdiction point more to add weight to its arguments on the discretionary ground than anything else; and I rather felt Mr Gabriel Moss QC for the Administrators tended to magnify beyond the significance which the GTA itself attributed to it.

By way of summary, the argument depends upon the application of the DTC between Germany and the UK which was signed on 30 December 2010 and took effect in the UK from 1 January 2011 in relation to withholding taxes. The DTC is closely modelled on the OECD’s Model Tax Convention (“MTC”). The Organisation for Economic Co-operation and Development (“OECD”) publishes official commentaries on the MTC and in considering a provision of a double-taxation convention that mirrors one in the MTC, it was not disputed that the English courts will have regard to the OECD commentaries on the equivalent provision in the MTC (see also Sun Life Assurance v Pearson [1987] STC 461, per Vinelott J at 510).

Article 28 of the DTC is headed “Assistance in the collection of Taxes”. Article 28(1) provides:

“The Contracting States shall lend assistance to each other in the collection of revenue claims…”

For this purpose, a “revenue claim” is defined by Article 28(2): “The term “revenue claim” as used in this Article means an amount owed in respect of taxes of every kind and description imposed on behalf of the Contracting States, a “Land” or of a political subdivision or local authority of a “Land” or a Contracting State, insofar as the taxation thereunder is not contrary to this Convention or any other instrument to which the Contracting States are parties, as well as interest, administrative penalties and costs of collection or conservancy related to such amount.”

Article 28(3) provides:

“When a revenue claim of a Contracting State is enforceable under the laws of that State and is owed by a person who, at that time, cannot under the laws of that State, prevent its collection, that revenue claim shall, at the request of the competent authority of that State, be accepted for purposes of collection by the competent authority of the other Contracting State. That revenue claim shall be collected by that other State in accordance with the provisions of its laws applicable to the enforcement and collection of its own taxes as if the revenue claim were a revenue claim of that other State.”

Then Article 28(6) provides:

“Proceedings with respect to the existence, validity or the amount of a revenue claim of a Contracting State shall not be brought before the courts or administrative bodies of the other Contracting State.”

It is Article 28(6) which the GTA has suggested exclusively reserves its GTA Claim to the German Courts. Mr Tom Smith QC (leading Mr Andrew Shaw) submitted that the Article is free-standing, and is not limited to the context of a disputed enforcement claim after a request for assistance. He submitted that it leads to an entirely sensible position, which is that an established ‘foreign’ revenue claim may be enforced in this jurisdiction with the assistance of HMRC; but in the case of any challenge, in whatever context (except in the case of express legislative exception), to the claim’s existence, validity or amount then the issue must be remitted to the ‘foreign’ country seeking to establish the claim. On that footing, the only erosion (otherwise in the context of insolvency) to the well-established (and almost universal) principle in Government of India v Taylor [1955] AC 91, that the courts of one country will not enforce the penal and revenue laws of another country, is that such enforcement is possible but only with the assistance of HMRC after prior request for it.

Mr Smith accepted, of course, that the provisions of Article 13(3) of Schedule 1 to the Cross-Border Insolvency Regulation 2013 (“the CBIR”) preclude the rejection of a proof of debt:

“solely on the ground that it is a claim by a foreign tax…authority but such a claim may be challenged…(b) on any other ground that a claim might be rejected on a proceeding under British insolvency law.”

However, he emphasised that there is nothing in Article 28(6) to provide or suggest any carve-out in respect of insolvency proceedings, and he submitted that the two provisions, Article 28(6) of the DTC and Article 13(3) of Schedule 1 to the CBIR are not opposed, but rather sit alongside each other, the former dealing with the prior stage of establishing the claim and the latter with the enforcement stage.

Accordingly, Mr Smith argued, the GTA Claim falls within the exclusive jurisdiction of the German Tax Court and the GTA Appeal must be stayed.

Against this, Mr Moss submitted that Article 28(6) is not a free-standing rule, but must be read in the context of Article 28 as a whole. Article 28 of the DTC introduces a self-contained exception to the common law rule of which Article 28(6) is an integral part. A request for assistance is the trigger for the regime in the article as a whole, not the mere existence of a revenue claim.He submitted that Article 28(6) merely confirms that, where collection of the taxes of a requesting state is requested or conservancy measures are undertaken, the determination of the substantive tax liability should be by the applicable courts or authorities of the state where the tax is owed: the rule under Article 28(6) is necessary because the determination of whether the tax is “enforceable” is a prerequisite to a request for assistance. The tax debtor cannot re-argue the substantive liability as a defence to collection activity undertaken by the competent authority of the requested state on behalf of the requesting state. Similarly, a tax debtor who does not contest the liability in the state where the tax is allegedly owed, cannot defend collection activity in the requested state by contesting the substantive liability to tax.

Mr Moss sought support in this context from the OECD Commentary to Article 27(6) (corresponding to Article 28(6) of the DTC), at paragraph 28, which confirms the application of the provision to cases where assistance is requested within Article 28:

“This paragraph ensures that any legal or administrative objection concerning the existence, validity or the amount of a revenue claim of the requesting State shall not be dealt with by the requested State’s courts and administrative bodies. Thus, no legal or administrative proceedings, such as a request for judicial review, shall be undertaken in the requested State with respect to these matters. The main purpose of this rule is to prevent administrative or judicial bodies of the requested State from being asked to decide matters which concern whether an amount, or part thereof, is owed under the internal law of the other State. States in which the paragraph may raise constitutional or legal difficulties may amend or omit it in the course of bilateral negotiations.”

Thus, according to Mr Moss, the fundamental difficulty with the suggestion is that the application of the entirety of Article 28, including Article 28(6), is confined to cases of assistance sought and provided by the competent authorities within the framework of the article, and there having been no request for assistance, Article 28(6) has never been engaged.

I tend to agree with Mr Moss that Article 28(6) should be read as confined to its context, and not to lay down any separate jurisdictional rule, even though it reflects and to that extent emphasises the well-established principle referred to above. But in any event, I do not accept Mr Smith’s argument that Article 13 of Schedule 1 to the CBIR and Article 28(6) can be read as he suggests so as to mandate the Administrators in effect to abandon the process of proof (and appeal from its rejection) in respect of an underlying foreign tax claim as soon as the validity or amount of the claim is questioned: it seems to me that his submission would subvert the process of proof, which extends to the determination subject to appeal of the validity of a claim as well as its enforcement. Thus, in my view, whatever may be the position where the CBIR does not apply, in an insolvency situation where it does apply Article 28(6) of the DTC is no impediment.

That leaves open, of course, the question of discretion: but it seems to me to make clear that as a matter of jurisdiction there would be no impediment to dealing with the GTA Appeal (and indeed all three appeals) in this jurisdiction.

Accordingly, I do not consider that the German Tax Courts have exclusive jurisdiction to determine the GTA Claim so as to require this court to stay the GTA Appeal: the decision whether to stay is discretionary.

On the issue of discretion, on which Mr Smith focused, the GTA’s principal submissions were as follows:

Against this, the Administrators’ principal submissions on the issue of discretion in relation to the GTA’s application may be summarised as follows:

In addition to the rival contentions adumbrated above on the issue of discretion a further matter was also argued between the parties. This was a question raised, by reference to evidence from German law experts, as to whether the German Tax Courts are in fact an available forum.

Neither party had obtained permission to adduce expert evidence. However, in Deutsche Bank Ag v Comune di Savona [2018] EWCA Civ 1740, Longmore LJ considered that CPR Part 35, which requires permission for any expert evidence to be adduced at trial, does not apply in the context of interlocutory applications (though he recommended that this should be reviewed). I have considered the arguments raised by the experts accordingly, though of course it is not appropriate to try to determine matters in contest between them before trial (and without cross-examination).

The questions raised, in summary, are as to (a) whether, under German law, the GTA had any power to issue Amended Tax Assessment Notices amending the original tax assessment after the commencement of a foreign insolvency proceeding; and (b) whether, following a foreign insolvency, it is competent for a German court to be seized of the substance of any disputed tax liability, or whether the German court would be confined to an adjudication as to the validity of the Amended Tax Assessment Notices.

In a report dated 18 October 2018 on behalf of the Administrators, Prof. Dr. Dietmar Gosch (“Prof. Gosch”), a Former Presiding Judge at the FFC, has expressed the view that the German court cannot proceed to a substantive determination of the reasons for a tax refund asserted in the Amended Tax Assessment Notices. That is because, in his opinion, there is no power to amend an original tax assessment following the commencement of a foreign insolvency proceeding; any tax claim must be proved in that proceeding; and, in consequence, the German court can only determine the (in)validity of the Amended Tax Assessment Notices.

Prof. Gosch has explained that in the purely domestic context, in his view, there is, in summary, a suspension of normal tax assessment powers, including of amendment, and instead a special process instituted that integrates with the German insolvency proceeding but which does not permit the creation of a separately enforceable tax debt. A disputed tax claim so notified can thus be appealed and adjudicated upon by the German court. However, in the circumstances of foreign insolvency proceedings, in his view, the assessment powers are again suspended, including the power to amend previous assessments, and instead, any tax claims are to be asserted in the foreign insolvency and dealt with according to the law of that proceeding (paragraph 40). In consequence, Prof. Gosch’s view is that any purported post-insolvency assessments, including amendments to previous assessments, are void, and the German court is bound to decide any challenge to any post-insolvency assessments in terms of assessing the validity of the assessments only.

Furthermore, according to Prof. Gosch, the Amended Tax Assessment Notices are invalid; and, unlike in the case of a German insolvency proceeding, which integrates a disputed notification of a tax liability into the German court system, there are no rules in the case of foreign insolvency which enable the German court to be seized of the substance of any disputed tax liability.

The GTA has filed and served a response on German law from Prof. Dr Heribert M. Anzinger (“Prof. Anzinger”), a Professor of Business Law and Tax Law at the University of Ulm. In his view, contrary to that of Prof. Gosch:

In the context of point (4) in the preceding paragraph, Prof. Anzinger also suggests that Prof. Gosch’s assertion to the contrary is founded upon the fact that the Administrators have so far only raised this formal issue in their objection to the Amended Tax Assessment Notices in Germany, and have not raised any substantive objections. But, in Prof. Anzinger’s view, the German Fiscal Court, once seized of the dispute as to the validity of the Notices, would seek to determine both formal and any other substantial challenges to tax assessments thus before it. The likelihood is that the Administrators would in practice be bound to raise their substantive objections to the GTA Claim (as they would be entitled to do by amendment) for fear of the German Fiscal Court determining the matter in the round: if the Administrators were to choose not to raise their substantive objections to the Claim in Germany they could not complain that the German court had not properly determined the matter. Further and in any event, according to Prof. Anzinger, the issue is unreal:

As I have previously indicated, I cannot at this stage properly determine which of these competing views as to the justiciability of the substantive matter in Germany is correct. In recognition of this, the Administrators have submitted that nevertheless, the Court can at least see (a) that it is not certain that the German court is an available forum for the resolution of the disputed issues in the GTA Appeal; and (b) even if it is, there is likely to be a disputed preliminary issuein any proceeding in Germany concerning the Amended Tax Assessment Notices, which will result in considerable additional delay. In such circumstances, the Administrators urge me to conclude, by reference to established principles, that the proper course is to refuse the GTA’s stay application.

Although the practical considerations raised by Prof. Anzinger seem to me to be quite persuasive, the doubts raised by Prof. Gosch have caused me concern: the availability of the alternative forum is a premise of any application for a stay to enable proceedings in that forum. However, the question really is whether the doubt is sufficient to compel the choice in favour of this jurisdiction, or simply another factor to be weighed; and I need to consider also whether some other means of addressing the point might be available, for example, by requiring undertakings from the GTA that if the German court cannot deal with the substantive matter they will withdraw their proof.

I do not feel able to determine the issue of availability, and whether the weight to be given to any doubt in that regard will be affected according to what would happen to the proceedings in this jurisdiction if Prof. Gosch is found to be right and the German Fiscal Court cannot determine the substance of the dispute.

In that connection, I understood Mr Smith on behalf of the GTA to accept in the course of the hearing that if the German Fiscal Court finds that the Amended Tax Assessment Notices are invalid, and that the GTA cannot validly issue a different form of notice (referred to in the evidence as an insolvency assessment notice), that is the end of the GTA Claim; but it may also be necessary to close out the possibility that, though valid in form, they cannot be adjudicated in Germany in substance; and the undertaking I have in mind would close off the possibility of the GTA then seeking to re-start the matter in this jurisdiction after an expensive, but in the event fruitless, expedition in Germany. In other words, the GTA should be put to their election and required to accept the consequences if the forum they have maintained is available turns out not to be so.

I shall, therefore, approach the matter on the footing that the German Fiscal Courts are to be treated as an available forum, but that one way or another, even if that is incorrect, a definitive result will be obtained in terms of the GTA Proof/Appeal if the matter is (in effect) remitted to Germany, either by adjudication or (if the German

The GTA entered its proof without reservation, in order that its claim should not be shut out by the hard bar date proposed by the CVA, and so that the GTA could then participate in the CVA which was at that time the intended means of distribution in the estate. The GTA has then taken the further voluntary step of appealing the rejection of its proof. Even accepting (as I do) that the GTA had no realistic alternative, this amounts to submission to the process mandated in this jurisdiction even though that meant running the risk of submitting to determination in this jurisdiction of a claim which by issuing its Amended Tax Assessment Notices it had earlier sought to have determined in Germany. Plainly the burden is on the GTA to establish good reason why it should not be held to that choice and why, more broadly, the process prescribed for, and usually especially appropriate for, the determination of claims in an insolvency process, should not be adopted.

Further, there are sound reasons why, if it were possible, all these issues, including the assessment of the value of DB’s contingent claims, could be finally determined in one process: but for the Later MFGUK Refund Claim, that would be possible in England, whereas on any view, the adjudication of all these matters cannot be undertaken in Germany.

My own estimation – and it is only that, since the time estimates offered were very tentative – is that even with the burden of matters of German law having to be proven by reference to extensive and contested expert evidence, it is probable that a full trial could be completed earlier in this jurisdiction, which has already demonstrated expedition in dealing with matters arising in the context of this important special administration. Likewise, I would expect a swift progress to the Court of Appeal and expeditious determination there.

By contrast, I have also been concerned by the very extensive period of time that it has taken the GTA to bring forward its claims and to respond to the Later MFGUK Refund Claim (first made in 2012 and not responded to until 2017). The transactions took place almost a decade ago. The GTA did not issue its four Amended Tax Assessment Notices until 8 December 2017 (after the notification of the CVA Proposal which was approved on 12 December 2017). The time that has elapsed before any proper adumbration of its various claims also encourages doubt as to the estimates that have been offered up on behalf of the GTA as an indication of the likely time before the matter is likely to be concluded in the German Fiscal Court, let alone on appeal.

I bear in mind also that the previous delays whilst the German authorities have investigated and assembled their case in response to ‘Cum/Ex schemes’, which have been adopted by many others in addition to MFGUK and which have resulted in substantial losses to the German Treasury, have had damaging results for the other creditors in MFGUK’s special administration. The emergence of the DB €127m Claim, consequentially upon an intimation of a claim against it by the GTA and only shortly before the CVA became effective, occasioned the lapse of that CVA proposal,

In addition, I am not wholly convinced that issues as to the systemic effect and application of provisions of German law will take centre stage. The fact that at least in the context of the GTA’s broader grounds there is likely to be necessary a detailed exegesis of the details of the complex sequence of transactions, their true purpose, their commerciality (or its lack), and the intentions of those concerned in them, which is very much within the ordinary experience and expertise of this Court, has much weighed with me.

It does seem to me that in the context of any necessary factual enquiry, this Court may be more used to the process, and many of the witnesses (as well as documentary evidence) will be here. I was not given very much information about the procedures and process of the German Fiscal Courts. I asked questions at the hearing about such matters as (a) whether the process is inquisitorial or adversarial (I was told it is inquisitorial); (b) whether there were procedures for the disclosure of documents (this remained unclear); (c) whether witnesses would be compellable (which I was told, on instructions, they probably would be if within the jurisdiction); (d) whether there would be oral evidence (I was told, again on instructions, there could be, with the judge undertaking any examination). But the overall impression I formed was that the German Fiscal Courts may be best equipped to deal with matters of law rather than fact: the work and usual focus of a specialist tax court may not suit it to the process of sorting out difficult factual issues, though Mr Smith made the fair point in reply that it is unrealistic to think that the German Fiscal Courts do not have the necessary experience given that the record shows that they have already dealt with two cases which are not altogether dissimilar to this.

It has struck me also that the fact that the FFC, in the only case on Cum/Ex schemes that has so far come before it on appeal, chose to focus on the particular detailed contractual arrangements and what it held to be the artificiality of the transactions involved, rather than on detailed provisions of German tax law or on systemic principle, may encourage lower courts (the German Fiscal Courts) to focus on an analysis of the particular facts and circumstances, at least in respect of certain aspects of the various claims.

It has in that context also weighed with me that such facts may indeed be more easily explored here given (a) the disclosure obligations in this jurisdiction, (b) the residence and thus compellability of witnesses based in this jurisdiction, and (c) the probable location of documentary evidence here. That may also have repercussions in terms of relative costs, though it is not easy to assess with any accuracy what would be the difference in cost terms according to whether or not a stay is granted. Refusing a stay may save costs in respect of the exploration of the factual issues, but it would be likely to increase the costs of the legal issues (because of the need for expert evidence of German law); and it would commit the Administrators (unless they sell the Later MFGUK Refund Claim) to litigation in two countries.

There are thus many imponderables; and there are risks both ways. I have therefore had considerable misgivings about granting any stay. Nevertheless, and subject to undertakings as previously mentioned (see paragraphs [98] to [100] above) three

The first such factor is the point at the forefront of Mr Smith’s (and indeed Mr Fisher’s) submissions and has particular weight in consequence of the fact that the Later MFGUK Refund Claims are to be adjudicated in Germany. If no stay is granted, broadly the same issues would fall to be considered by the court here and the court there at (again speaking broadly) the same time and between the same parties. There is an obvious risk of inconsistent, indeed conflicting, judgments.

That is always capable of amounting to a very strong reason for granting a stay, as the cases I have referred to in paragraph [61] above show and emphasise. Thus, in Curtis v Lockheed Martin UK Holdings Ltd, it was because the grant of a stay would not, in circumstances where the claimant was not a party to the foreign proceedings and would not be bound by their result, remove the risk of inconsistent findings that such stay was refused; but the potential weight of the possibility of inconsistent findings (and a fortiori decisions) was expressly recognised. And in Prifti (again see paragraph [61] above) the fact that there were concurrent proceedings in Spain in which the Spanish court would be required to determine also the principal issue in the English proceedings, being whether a ‘pre-existing conditions’ clause had been validly incorporated into the parties’ contractual arrangements, so that there was a risk of inconsistent determinations on a fundamental issue, appears to have been the decisive factor in favour of a stay (although the judge considered also that the stay would be unlikely to cause material prejudice which could not be compensated for by an award of interest).

The desirability of seeking to remove that risk where possible without undue prejudice is not only a matter of judicial consistency. I agree also with the GTA and DB that there is a real possibility, perhaps likelihood, that if the two sets of proceedings go forward to adjudication at first instance, then whatever the sequence, practical conundrums will develop. If the proceedings here are determined first, but before any appeal to the Court of Appeal (of which, despite the fact that questions of German law will be treated as issues of fact in England, there must at least be a possibility given the complexity of the issues and the amounts at stake) a question is likely to arise whether any such appeal should be deferred pending any anticipated pronouncement of any contested issue of German law by the German Fiscal Court. If the German proceedings are determined first, issues will arise as to the extent that the findings bind the Court here, especially if an appeal is then pending in Germany to the FFC. One way or another, in other words, the difficulties may change but they will not disappear: only consolidation or concurrent hearing in one trial of the issues of German law, and a consolidated or concurrent appeal there would avoid these difficulties (though it cannot achieve the final determination of the issues relating to DB).

The second factor is this. It seems to me that, despite my hunch that there will also be considerable factual enquiry, and a factual determination of the particular circumstances may determine the result (as in the case in 2014 before the FFC), the legal issues at stake are not only plainly matters of German law, but controversial and complex issues of statutory construction of systemic importance and substantial public interest in terms of the legitimate interests of the public in the protection of its taxation system from what are alleged to be colourable schemes.

As it seems to me, the “potential disaster from a legal point of view”, as in The El Amria [1981] 2 Ll. Rep. 119 (at 128) Brandon LJ (as he then was) described the risk of inconsistent decisions in concurrent proceedings in different jurisdictions, is the more acute when in one of the jurisdictions the issue is a systemic one, or may be decided in a manner which has systemic consequences. Especially in such a context, there is a preference for a case to be heard by the courts of the country whose law applies: see VTB Capital v Nutritek International [2013] 2 AC 337 at [46] per Lord Mance:

“it is generally preferable, other things being equal, that a case should be tried in a country whose law applies. However, this factor is of particular force if issues of law are likely to be important and if there is evidence of relevant differences in the legal principles or rules applicable to such issues in the two countries in contention as the appropriate forum.” (Footnote: ⁶). See also Dicey, Morris & Collins on the Conflict of Laws (15^th Edn, 2018) at 12-034

Also in that context, even if the factual centre of gravity may be London, the jurisdiction likely to be most affected by the result is Germany: and even if the US approach of ‘interest analysis’ is not determinative in this jurisdiction it does not seem to me to be an impermissible consideration.

A third factor, relevant to the question of prejudice which is always an important one, is that it is an unusual feature of this case that in fact, the Later MFGUK Refund Claim is to be determined in Germany in any event, and the administration cannot finally be brought to an end until that and other matters are concluded.

Thus, whether or not a stay is granted, it does appear that MFGUK’s administration is likely to be kept open for another seven, perhaps up to twelve, years, partly because if the Later MFGUK Refund Claims, and partly because of another claim to recover European Union tax based on article 63 of the European Treaty. That is, in the aftermath of the failure of the CVA Proposal, the Administrators’ own estimate in their most recent report to creditors. Mr Moss’s response on their behalf to the effect that the Later MFGUK Refund claims may be assigned and thereby an earlier termination achieved is possible; but it is not what the Administrators have appeared primarily to have envisaged.

My concerns as to the potential prejudice in consequence of delay has also been attenuated by the consideration that, unusually, creditors have already received distributions amounting to 90p in the £; and all client money and asset entitlements have been dealt with.

Taking careful account of the caveat expressed by Lord Bingham in the Reichhold case (see paragraph [58] above), and of the burden on the GTA in light of its proof and appeal and the circumstances outlined above, I consider that these three factors in combination constitute a sufficiently “rare and compelling” reason for granting the stay sought by the GTA, provided that the German Fiscal Court are an available forum in which to determine the substance of the disputes.

As to that, I have made clear that I shall require undertakings as described in paragraphs [98] to [100] above (but which will need refinement) so that the stay does not result in an expensive detour to the same ultimate destination. Any stay will be conditional accordingly. I am also concerned about the very considerable delays thus far, primarily on the part of the GTA in formulating and after some time pursuing what it has maintained is a matter of public concern and import to the German treasury. I am also concerned lest this case become enmeshed in others which are presently at an earlier stage in the GTA’s assessment of them. I have in mind to require assurances, possibly by way of undertakings, to seek to ensure, insofar as the parties’ best endeavours can secure it, resolution of both the GTA Claim and the Later MFGUK Refund Claim as expeditiously as possible. That seems to me necessary in order to safeguard this jurisdictions’ insolvency processes and for the protection of the interests of the body of creditors as a whole.

DB submits that on that basis, if the GTA Proof/Appeal is to be stayed in favour of resolution of the various German law issues raised by the GTA Claim in Germany, it is only logical that its own appeals (the DB Mirror Proof/Appeal and the DB €127m Proof/Appeal) should also be stayed. That is essentially because (so DB submits) the same substantive German law issues arise for determination, the same “forum factors” favouring Germany for resolution of the substantive German law issues will apply, and there will be a high risk of inconsistent decisions if the English court were to proceed to determine its appeals before, and without the guidance of, the German Fiscal Courts.

It is relevant to clarify, however, that DB accepts that there are no proceedings against it in Germany by the GTA or MFGUK’s clients, and it has not evidenced any other form of claim that has in fact been made against it. It follows that the stay DB seeks is to that extent open-ended, and depends (presently at least) not on any formulated and actual claim against it but on claims by and against the GTA to which it is not a party and over which it has no control.

The basis of the DB Mirror Proof/Appeal is that DB anticipates that the GTA may, in parallel and addition to its GTA Claim, seek to hold DB liable in respect of its role in acting for MFGUK as seller and in submitting the withholding tax reclaims on its behalf. Liability is said to arise pursuant to clauses 9 and 10 of the terms of admission to the EFP (which are governed by German law).

In paragraph 68 of Voigt 1, Ms Voigt has stated that the GTA “is still considering its position in relation to Deutsche”. However, she has explained that under German law, any such claim as the GTA might make against DB would be

“separate and additional to its right to pursue MFGUK for the sums sought under the Amended Tax Assessment Notices (and therefore also the Proof in these proceedings), subject potentially to questions of avoiding ultimate ‘double recovery’”.

DB’s Mirror Proof was rejected by the Administrators on 8 February 2018. The reasons given for rejection were, in summary:

Both DB and the GTA, though otherwise adversaries in the matter, contend that, save for the issue as to ‘double proof’ (see (4) in the preceding paragraph) as to which the GTA takes no position, many of the same issues of German law which arise in respect of the DB Mirror Proof/Appeal also arise in respect of its own GTA Claim and Proof/Appeal, and in the Later MFGUK Refund Claim; and both contend on that basis that all such issues should be determined in Germany. As further elaborated below, the Administrators contend that this is beside the point, because it naturally follows from the application of the ‘rule against double proof’ that it will be dispositive and that on that basis no adjudication of the German law issues will be required to determine the DB Mirror Appeal.

DB lodged the DB€127m Proof on 12 January 2018. As will already be apparent, it is based on DB’s perception of its contingent exposure to claims against it by MFGUK’s clients in respect of its conduct as custodian bank acting for DB as seller in a raft of Cum/Ex transactions. It is additional to the DB Mirror Claim which relates to transactions in which DB was custodian bank for MFGUK when it was the buyer in Cum/Ex transactions which the GTA has impugned.

In more detail, DB’s concern is based on a decision of the Fiscal Court in Hesse dated 10 March 2017 (which I have referred to earlier in paragraphs [46] and [47] above). On the basis of that decision, DB now claims that either:

DB asserts that such potential liabilities, whether to the GTA or to MFGUK’s clients, means that it is entitled to an indemnity from MFGUK on the basis:

The Administrators rejected the DB€127m Proof for the following reasons, in summary:

In addition to the above reasons, the Administrators formally estimated the DB Indemnity Claim to have a nil value.

Both the GTA and DB contend that the issues raised (or likely to be raised) in the context of both DB Proofs/Appeals are, or predominantly concern, issues of German law which are common to, and will fall to be determined in the course of, the GTA Claim.

Conversely, and as further elaborated later, the Administrators contend that there is no substantial connection between the DB€127m Proof/Appeal and the GTA Claim.

Thus, although DB’s application to stay both its Mirror Proof/Appeal and its €127m Proof/Appeal are based on the contention that they both raise issues of German law common to the GTA Claim and should be disposed of in Germany, the Administrators object in each case, but on different grounds, though such grounds are based on the application of English law. The stays sought by DB in respect of its Mirror Proof/Appeal and its €127m Proof/Appeal respectively require separate analysis accordingly.

DB has identified the following issues of German law as being fundamental to its Mirror Proof/Appeal:

The Administrators raise disputes in relation to all of these DB Issues, and in addition rely upon the fact that no payments have as yet been made by DB to the GTA for which it could have any accrued right of indemnity as against MFGUK. However, it appears to be common ground that DB Issue 1 is the same issue as arises in the

Although it is contingent on the answer to DB Issue 1 being in the affirmative, DB Issue 2 does not otherwise appear to arise in the context of the GTA Claim (which is against MFGUK). However, although not presently the subject of any subsisting process in Germany, according to Mr Craig Montgomery’s first witness statement on behalf of DB dated 19 September 2018, the GTA has indicated that it intends to take steps to claim those refunds from DB, such that DB considers that proceedings are highly likely to be commenced in the German Fiscal court addressing both Issue 1 and Issue 2 for that purpose. On that basis, DB contends that, as with Issue 1 in the context of the GTA Stay Application, these questions are better determined in Germany so as to avoid the risk of inconsistent judgments, as well as for the other reasons relied upon in support of the GTA Stay Application.

DB Issue 3 (i.e. MFGUK’s indemnity obligation) appears to be contested solely on the basis that MFGUK denies that the underlying GTA Claim is valid. On that basis, DB submits that if a stay of the DB Mirror Proof/Appeal is otherwise considered appropriate, there is no real scope to consider determining Issue 3 in advance of the outcome of the German proceedings. DB does not understand the JSAs to deny that there is a right to indemnity or compensation on one or other of the bases relied upon by DB if the underlying GTA Claim is valid.

I did not understand the Administrators to dispute that, if (as I have determined) the GTA Claim is to proceed in Germany, there would be logic in the three issues thus identified as being fundamental to the DB Proof/Appeal also being adjudicated there, but only if (and this is the nub) such claims would be admissible to proof in this jurisdiction if established in Germany.

The nub of the Administrators’ objection to any stay of the DB Mirror Claim is their submission that such claims, even if established under German law, are simply not admissible to proof. They submit that the admission of such proofs is subject to mandatory English insolvency rules which preclude or bar them, and in particular, at least for so long as the GTA also maintains its Claim, the ‘rule against double proof’.

The Administrators’ case in this regard relies on the fact that the DB Mirror Appeal is dependent upon the validity of the GTA Proof/Appeal, being a contingent claim against MFGUK under an alleged indemnity that is dependent (as to its quantum) upon the GTA successfully claiming against DB for, in substance, the same debt as the GTA has proved for against MFGUK.

The Administrators submit that there are two logical possibilities:

The Administrators submit that it naturally follows from the application of the ‘rule against double proof’ that the DB Mirror Appeal is bound to be dismissed. If the GTA Claim fails, DB has itself no claim; if it is assumed to succeed, DB’s Proof is inadmissible as being a second proof in respect of substantially the same ‘debt’ as that claimed by GTA. Whatever may be (or envisaged to be) the result in Germany, therefore, there is no justification for, or benefit to, any stay of that appeal, and the estate will benefit from the dismissal of the DB Mirror Appeal, and the certainty which would flow from it.

In answer, and notwithstanding paragraph 68 of Voigt 1 (quoted in paragraph [127] above), DB does not dispute that the GTA Proof/Appeal and its Mirror Proof/Appeal for €48m are competing claims which, in substance, are claims for payment of the same debt twice over in the required sense such that dividends cannot be permitted to be paid in respect of both. DB accepts that the existence of DB’s Mirror claim is dependent on the existence of a liability owed by MFGUK to the GTA, and DB accepts that it would produce injustice to other creditors if both qualified simultaneously for dividend. However, DB disputes that the ‘rule against double proof’ is such as to preclude or debar its proof at this stage. DB relies on two main interrelated arguments.

First, Mr Fisher submitted, the so-called ‘rule against double proof’ in English law is a misnomer. The true rule is not against double proof but against double dividend. It operates to bar the payment of two dividends in respect of what is in substance the same debt, rather than the presentation of two proofs for the same or substantially the same debt. This argument concerns a legal question as to the true nature of the ‘rule against double proof’ in English law.

Secondly, Mr Fisher submitted, this is a case where there is, on the facts, real doubt as to which proof will ultimately have the superior right to be admitted for dividend, so that the Administrators’ submission that the DB Mirror Proof/Appeal should be treated now as barred by the rule against double proof is premature (and potentially ill-founded). That is because, even though the GTA did submit an appeal against the rejection of its proof of debt, there remains a real prospect that the DB Mirror Proof/Appeal may be the relevant proof which should be admitted in the MFGUK estate. This argument concerns the factual question as to whether there is any outstanding uncertainty as to whether in the end there will be any issue of ‘double proof’, or whether, as a matter of fact, it may be that only one claim will remain to prove, and that will be on the part of DB, and not the GTA.

Mr Fisher particularly relied in respect of both arguments on the decision of the Court of Appeal in Barclays Bank v TOSG Trust Fund Ltd [1984] AC 626 (“Barclays v TOSG”, which went to the House of Lords, where it was held that no issue of double proof arose on true construction of the contractual arrangements (see 675D-E) and was thus decided on a different basis, but without disapproval of or adverse comment on the Court of Appeal’s analysis of the rule).

In Barclays v TOSG, the Court of Appeal unanimously rejected the submission that the rule is always to be applied at the date of the commencement of the insolvency (in that case, a liquidation) or at any stage prior to the point when a dividend is about to be paid: only at that stage and in the circumstances by then apparent does there have to be a decision as to whether there is in reality a danger of double recovery and, if so, which of two or more claims should be given priority.

Oliver LJ (as he then was) squarely re-characterised the rule as in reality a ‘rule against double dividend’. In doing so, he focused especially on the crucial question as to when the rule is triggered, and answered it by reference to the date when he considered the mischief of the rule was designed to prevent eventuates. He considered the matter at 636C and 637H to 638B as follows:

“No doubt it can be predicted at the commencement of the liquidation that a case for the application of the rule may arise or that it can never arise, but it may well be impossible to determine at that stage whether it will in fact.

…

Now if, as in my judgment these cases show, the true rule is that there are not to be two dividends in respect of what is in substance the same debt, I can see no logical justification for seeking to fix the position at the commencement of the insolvency. One has, as it seems to me, to look at the position at the point at which the dividend is actually about to be paid and to ask the question then whether two payments are being sought for the liability which, if the company were solvent, could be discharged as regards both claimants by one payment.”

Slade LJ’s judgment in the same case was to the like effect. At 659G to 660B, citing Mellish LJ’s judgment in In re Oriental in support, Slade LJ said this:

“…Mellish LJ had likewise made it plain that the rule is directed against the payment of more than one dividend in respect of the same debt, rather than against presentation of more than one proof. In many cases, such as the present, where more than one proof has been presented, one may find what was sometimes described in argument as a “potential double proof situation”, which can only be finally resolved at a latter stage, having regard to the facts subsisting at the time when a dividend is about to be paid (for example, having full regard to the arrangements made pursuant to the assignment agreement in the present case). The purpose of the rule is, of course, to ensure pari passu distribution of the assets comprised in the estate of an insolvent in pro rata discharge of his liabilities. The payment of more than one dividend of what is in substance the same debt would give the relevant proving creditors a share of the available assets larger than the share properly attributable to the debt in question.

Difficulty may well arise in determining whether, in any given case, two proofs are in respect of what is in substance the same debt. Though various broad tests have been canvassed by both Bar and Bench in argument in this case, I have, for my own part, found none of them wholly satisfactory. The question can, I think, only be determined by reference to the particular facts of the case before the court, bearing in mind that it is the substance of the relevant liability, rather than the form, on which attention must be concentrated.”

Mr Fisher also relied on the Supreme Court decision in In reKaupthing Singer & Friedlander Ltd (in administration) (No 2) [2012] 1 AC 804 (“Kaupthing”), which he submitted in effect approved the Court of Appeal’s decision in Barclays v TOSG on

Relying on those two cases, Mr Fisher presented the matter simply, depicting the issue as “a relatively discrete matter of timing as a matter of English law”, where the facts demanded admission of both proofs and a later determination as to which should qualify for any future distributions.

DB maintains that this is a case where there is real doubt as to which proof will ultimately have the superior right to be admitted for dividend, and that the Administrators’ submission that the DB Mirror Proof/Appeal should be treated now as barred by the rule against double proof is premature (and potentially ill-founded). Even if it may appear that DB’s exposure is secondary, and even though the GTA did submit an appeal against the rejection of its proof of debt, there remains a real prospect that the DB Mirror Proof/Appeal may be the relevant proof which should be admitted in the MFGUK estate.

For example, if the GTA pursues DB in Germany and if, which is denied, DB is found to be liable, DB may be forced to make a payment in full to the GTA in respect of the tax reclaims that the GTA is pursuing. In such an event, DB would contend that the GTA proof should be rejected and that the DB Mirror Proof/Appeal would be the only claim that ought to be admitted to proof for dividend in the MFGUK estate. In other words, DB would in those circumstances have the superior right of proof, which should be admitted: see Lord Hoffmann’s comments in Secretary of State for Trade and Industry v Frid [2004] 2 AC 506 at [13] on In re Fenton (No1) [1931] 1 Ch 85; and Barclays Bank v TOSG at 643.

Mr Fisher submitted that the DB Proof/Appeal cannot be rejected on the basis of the ‘rule against double proof’ in such circumstances: (a) the ‘rule’, properly interpreted, does not require a decision until a distribution is to take place and it is uncertain to which of them it should be paid and (b) that uncertainty cannot presently be resolved and the determination between them should be stayed until after the determination of the GTA Claim in Germany. Only thereafter, he submitted, would “the nature of the rights arising as matter of German law and the way in which they interplay and the point at which [DB] would be allowed as a matter of German law to make its claim” become clear; and, he urged, “those are secondary questions that we need not get into at this time.”

Mr Moss rejected this approach and widened the debate. He depicted Barclays v TOSG as a variation and extension of the basic ‘rule against double proof’ to cater for cases where there was remaining doubt as to whether there was indeed the same or substantially the same debt, and he referred me to a number of authorities in relation to the parties’ competing contentions as to (a) the true nature of the rule and (b) the time at which it is to be applied in the course of an insolvency process.

Starting with the two authorities on which Mr Fisher concentrated, Mr Moss pointed out that in Kaupthing, Lord Walker also (at [12]) quoted Lord Hoffmann’s examination of further case law and summary of the position in Secretary of State for Trade and Industry v Frid as follows:

“In re Fenton; Ex p Fenton Textile Association Ltd [1931] 1 Ch 85 was another case of a surety under a pre-insolvency guarantee, but this time he had not actually paid. Nor could he pay, because he was bankrupt and his assets had vested in his trustee. The creditor was still owed the money and entitled to prove in the liquidation. The Court of Appeal held, first, that one could not have more than one proof in respect of the same debt (‘the rule against double proof’); otherwise, if there had been, say, four guarantors, there could have been five people receiving dividends on the same debt. Secondly, the Court of Appeal said that until the creditor had been paid, he had the superior right of proof and a right of proof by a surety was excluded…”

Those last words gave obvious comfort to Mr Moss, who focused much more attention on In re Fenton, a case primarily concerned with the issue whether a set-off

Lord Harnworth MR agreed in the result, but with the caveat (see page 110) that he did not mean to foreclose the possibility that the claimant might in the future establish a provable debt “in some events”; and Romer LJ, who focused more on the ‘rule against double proof’, considered In re Oriental Commercial Bank to be “direct authority against such a proof being allowed” (see page 119) and stated this (at pages 118 to 119):

“Should the surety subsequently pay off the principal creditor before the latter has lodged a proof, he would undoubtedly be able to prove in the bankruptcy, and if he paid the principal creditor off after the latter has lodged a proof, the dividends in respect of such proof would be made available for the surety. But I cannot agree that a surety who has not paid off the principal creditor can prove in the bankruptcy of the principal debtor so as to share in the distribution of his assets unless the principal creditor has renounced in some way his right to lodge a proof himself while preserving, of course, his rights against the surety. To allow such a sharing in the assets would be to subject the assets to two claims in respect of the same debt, and this is contrary to the well-established rule in bankruptcy against double proof.”

These citations may appear to suggest a tension between the approach of the Court of Appeal in Re Fenton and that of the Court of Appeal in Barclays v TSOG and the House of Lords in Kaupthing. Yet Re Fenton was cited in each without express or apparent disapproval; and, as appears above, it was cited with apparent approval by Lord Hoffmann, who, incidentally, had represented the unsuccessful appellant in the Barclays v TSOG (Footnote: ⁸), in Secretary of State v Frid. That begs the question whether there is any real dichotomy between the cases.

Mr Moss submitted initially that there is, and that to the extent that any dicta in the

Mr Moss cited further authority in his support as regards to the purity (and correct characterisation) of the ‘rule against double proof’ in cases of that nature (including, in addition to In re Oriental Commercial Bank and re Fenton, the decision of the House of Lords in Deering and Others v The Governor and Company of the Bank of Ireland (1886) 12 App. Cas. 20, per Lord Halsbury LC at 28, and of Vaughan Williams J in In re Sass, ex parte National Provincial Bank of England [1896] 2 QB

Mr Moss submitted that there should be no prejudice occasioned to DB by adopting that cut-off date: for DB would have a right as against the GTA which it could exercise by way of subrogation to the GTA’s right of proof: if, in the circumstance envisaged by DB as justifying its proof, the GTA looks to DB in Germany and achieves payment from DB (see paragraph [147] above), then in effect, the GTA would hold any dividends that they obtain in the future out of the estate on trust for DB (as indeed Vaughan Williams J considered in In re Sass at page 15).

In a rapid-fire but concise and well considered reply, Mr Fisher sought to rebut the suggestion of a different approach according to whether or not there was uncertainty at the date of proof as to whether competing claims arose out of substantially the same debt, or as to the priority between them: he submitted, overall, that even in the ‘plain vanilla’ case of surety and principal debtor, the rule took effect, not at the date of lodgement of proof, but at the date of distribution.

As to the cases relied on by Mr Moss, he submitted that particular care had to be taken to look at the factual circumstances in which each had been decided. He pointed out that in Re Oriental Commercial Bank, payments by way of dividend had already been made to the principal creditor (Agra Bank) at the point when the surety lodged its proof, so that there was an immediate prospect of double dividend on the proof admitted. As to Deering v Bank of Ireland, it was not possible to tell whether or not a payment had yet been made in respect of the principal debt, but (he submitted) that

More generally, Mr Fisher submitted that the lodging of a proof is merely a procedural step, which of itself has no substantive significance, and which may be withdrawn, rejected or expunged. The admission of proof and the calculation of dividend are likewise procedural matters for the liquidator or trustee. Mr Fisher drew my attention in this context again to Andrews and Millett’s ‘Law of Guarantees’ which at page 542 sets out the authors’ view that

“…it is unsatisfactory that the right of the surety should be dependent upon something as arbitrary as the date on which the creditor lodges his proof. The mischief of the rule against the double proof is to prevent a doubling-up of dividends, and the fact that there are two proofs for the same debt is something that can be dealt with by the liquidator or trustee when he calculates and pays the dividend, making unnecessary the application of any rule of law at any earlier stage.”

Notwithstanding his very thorough analysis of the principles and authorities on the

I recognise the call and attraction of caution. The ‘rule against double proof’ is and always has been invariably accepted to be an overarching principle of insolvency law even though there is no specific statutory provision to that effect, but the various authorities cited (almost all of which are binding upon me as decisions of the Court of Appeal or higher) reveal differences in its formulation and characterisation. I recognise also that there are undoubtedly differences of academic opinion, including as to which interpretation of the rule should presently be treated as having the preponderance of judicial and academic support.

However, and with diffidence, I do wonder whether some of these differences may stem from assimilating, or perhaps confusing, the ultimate objective or principle of the rule with the means whereby it is achieved.

As noted by Robert Walker J (as he then was) in Re Polly Peck International [1996] 1 BCLC (which the parties did not take me to, but which was included in the bundle of authorities and which I found helpful), the ‘rule’ is based on fairness, and depends on substance and not form. The principle of the rule, as stated by Mellish LJ in Re Oriental and Commercial Bank,is perfectly clear: that “an insolvent estate ought not to pay two dividends in respect of the same debt”, and (broadened a little to capture what is in substance the same debt, even if not in form) that:

“there is only to be one dividend in respect of what is in substance the same debt, although there may be two separate contracts”.

Even with that broadening of the meaning of “same debt” to include debts which are in substance the same, the rule can only, or only ordinarily, apply to a situation which actually is, or is analogous to, that of suretyship. Where the relationship between two potential claimants in the insolvency is one of suretyship or plainly analogous in its substantive effect (such as in the old cases on negotiable instruments considered in Re Oriental Commercial Bank) there is no doubt that a dividend can only be paid to one of the two claimants (the principal creditor C, rather than the surety S), at least unless and until S has paid C in full, whereupon S may assert C’s right.

In such a situation, the paradigm situation, the obvious and immediate way of safely securing the purpose or principle of the rule is to preclude any proof by S. That, as I read the cases (such as Re Fenton), is what the court has always done. The policy of the rule and the practice of the Court has been to determine the matter at the point when S lodges its proof. (Footnote: ⁹) There is no reason or justification for deferring the decision; on the contrary, the interests of all concerned in determining admissible proofs, which may for example affect the prospect as well as calculation of intermediate distributions, are better served by a clear rule applied at the date of the lodgement of proof.

However, in situations not so clearly constituting or equivalent to a relationship of principal creditor and surety, there is greater difficulty, which may only be resolved at a stage when the true relationship and priority between the competing claimants (in a similar position to C and S) can be determined. The Court may have to wait and see, although even then the game is over once a distribution is made, since on no account can two distributions be made in respect of what is in substance the same debt.

That, to my mind, may explain authorities such as Barclays v TOSG, which in a sense are examples of the court extending the ‘pure’ rule to capture other circumstances where the same basic unfairness may be revealed by further enquiry (of multiple dividends on what on analysis transpires to be substantially the same debt). In such circumstances, the “close jurisprudential analysis of the persons by whom and to whom the duties are owed” (as Oliver LJ put it in Barclays v TOSG at 636D) is required.

My provisional view is, therefore, that the authorities are not inconsistent: they address two different situations, one being or being analogous to the paradigm, the other being less obvious but potentially productive of the same substantive unfairness. In the first, paradigm, type of case, the authorities all stipulate rejection of the proof; in the other, they resort to the underlying rationale of the rule to permit deferral to the point when the risk of double dividend actually eventuates.

On that basis, the real question in this case is whether the relevant relationship between DB and the GTA with respect to the DB Mirror Claim is or is not analogous to the paradigm (the first type of case as I have described it above); or whether it is a case of the second type.

Two issues then require consideration: (a) whether the claims arise in respect of substantially the “same debt”; and if so (b) what the priority between the claims is as against the principal debtor (and in that context, what would be the effect if the GTA is entitled to recourse against DB, and satisfies its entire claim by such recourse).

Although the relationship between DB and the GTA is apparently governed by German law, I cannot presently see any reason (and none was suggested to me) for departing from DB’s own depiction of the two claims in its skeleton argument as referred to in paragraph [147] above. At the risk of repetition, there is no dispute that (as their description reflects) they are competing claims which, in substance, are claims for payment of the same debt twice over in the required sense such that dividends cannot be permitted to be paid in respect of both. DB accepts that the existence of DB’s Mirror claim is dependent on the existence of a liability owed by MFGUK to the GTA, and DB accepts that it would produce injustice to other creditors if both qualified simultaneously for dividend. Accordingly, in my view, the difficulty of determining whether two proofs relate to the substantially the same debt which arose (at least on the Court of Appeal’s view of the facts) in Barclays v TOSG and justified postponement of the decision which should qualify for distributions does not arise on that basis here. The concession removes the need to examine German law in relation to whether, for the purposes of the ‘rule’, the two claims can be said to be based on substantially the same ‘debt’: DB has very properly accepted they can and should be.

More difficult, to my mind, is the second question, and in particular as to the significance of the fact that the GTA is planning (according to the evidence of Jochen Von Berghes in paragraph 29 of his first witness statement dated 1 March 2018

The possibility on that basis, or because in the event the GTA is able to and does recover an amount equal to the entirety of its claim, of the GTA Proof/Appeal being withdrawn or rejected, leaving DB as the only claimant as against MFGUK without competition from the GTA, has caused me more than pause for thought as to whether, therefore, this is, after all, a case such as Barclays v TOSG where it cannot presently safely be determined which (if any) of two claims should be permitted to proceed by way of proof here.

My provisional view is that the answer to this also is provided by Re Fenton, and by reminding oneself that all that is presently to be determined is whether the stay sought by DB should be granted. In Re Fenton, the possibility of the proof which was held to be inadmissible subsequently becoming admissible upon the preferred proof being withdrawn was expressly contemplated. The answer given was that in that event either or both (a) the previously debarred proof could be re-submitted and/or (b) any distribution attributable to the previously preferred proof would become attributable and payable to the previously debarred proof (if otherwise valid): see the passages in Re Fenton quoted in paragraphs [161] and [162] above.

In any event, and whilst refraining from a definitive decision that the ‘rule’ applies so as to determine the DB Mirror Proof/Appeal, it is plain to me that it is certainly arguable that it does. In short, therefore, I decline to grant the stay sought by DB in respect of its Mirror Proof/Appeal. Whether, and if so in what form, that appeal should now proceed will have to be addressed further. Counsel are invited to consider this and the most efficient way of proceeding.

The DB €127m Proof/Appeal reflects, and seeks indemnity from MFGUK in respect of, all potential claims which DB has identified may be made against it in respect of the matters in question whether by the GTA or by ‘purchasing clients’ (being MFGUK’s former clients acting as alleged short buyers from MFGUK as short seller in Cum/Ex transactions). It was the sudden and late emergence of the DB €127m claim and proof that prevented the CVA proposed to bring an end to this long-running administration from taking effect.

I have described the basis of the DB €127m Proof/Appeal and the reasons for its rejection put forward by the Administrators in paragraphs [130] to [134] above. I turn

DB’s case is simply put: it contends that the DB €127m Proof/Appeal raises or depends in substance on the same issues of German law as does the GTA Claim and Proof/Appeal, and these issues are far more appropriately dealt with in Germany, both because they are matters of German law of systemic importance, and because of the real risk of inconsistent judgments given the pendency in the German Fiscal Court of the Later MFGUK Refund Claims and also (now that I have granted a stay) the adjudication of the GTA Claim in Germany. DB relies on all the cases previously considered in the context of the GTA Claim.

The Administrators cannot rely on the ‘rule against double proof’ in the context of the DB €127m Proof/Appeal: for that proof is based on a claim for indemnity in respect of claims that may be brought against it by clients of MFGUK and there is neither any risk of double proof and distribution nor anything akin to a surety situation: in other words, there is no question of its claim arising out of ‘substantially the same’ debt as another proving entity.

The questions to be addressed, therefore, are

As to (1) in paragraph [190] above, I do not think it is necessary, in view of my previous description of the issues in the GTA Claim and having regard to the length of this judgment already, to rehearse the common elements and issues of German law relied on by DB. Instead, I think it is more helpful to identify the differences relied on by the Administrators, and DB’s explanation why they are not substantial.

Before that, however, it is important to recall at the outset that the claims, the subject of the DB €127m Proof/Appeal, are claims by DB for an indemnity from MFGUK in respect of a contingent liability which it anticipates is liable to be triggered by claims by the GTA or by MFGUK’s customers (‘purchasing clients’) none of which has yet

As to the transactions said to give rise to the issues, the central point of difference relied on by the Administrators is that whilst in the GTA Claim, MFGUK is said to have acted as buyer against a short seller, in the DB €127m Proof/Appeal, MFGUK is said to have acted as seller and not as buyer. Thus, the Administrators submit, a key factual issue in relation to the GTA Claim and the DB Mirror Claim, which is whether or not MFGUK did act as a short buyer to a short seller in contrived Cum/Ex transactions, does not arise in relation to the anticipated claim against DB for which it seeks indemnity. Nor accordingly does the central and difficult German law question identified by the GTA as to whether a Cum/Ex purchaser is the economic owner of the shares the subject of the Cum/Ex purchase transaction and thereby entitled, or to be treated as being entitled, by the German tax authorities to any form of German withholding tax relief.

DB’s response is that this approach ignores the fact that the attack which DB fears and envisages, and which is its focus in its DB €127m Proof/Appeal, would be likely to come, not from the GTA, but from MFGUK’s customers (the ‘purchaser clients’ as defined in the evidence) who were buyers from it and were thus in the same position as MFGUK in the transactions, the subject of the GTA’s Claim. In his second witness statement on behalf of DB dated 21 December 2018 (“Montgomery 2”) Mr Craig Montgomery (“Mr Montgomery”, a partner at Freshfields Bruckhaus Deringer LLP, DB’s solicitors in this matter), provided (in paragraphs 21 to 22 and 29) this explanation by way of clarification:

“…DB does not contend that the relevant overlap was because MFGUK acted as seller in respect of both sets of transactions. Rather, DB contends that the principal relevant factual overlap underlying the two claims, and giving rise to an overlap or similarity in issues, is because the claims potentially arise in circumstances where MFGUK acted as cum/ex-buyer in the GTA appeal, and MFGUK’s former clients acted as cum/exbuyers in the €127m appeal.

As cum/ex buyers, MFGUK (in the context of the GTA appeal) and MFGUK’s former clients (in the context of the €127m appeal) both received dividend compensation payments in the amount of the net dividend. I am informed by my German colleagues that, based on the GTA’s current practice, it is to be expected that any tax refunds in relation to the withholding tax imposed by the German issuers will be or will have been rejected or reclaimed by the GTA. It is this underlying tax refund, and the GTA’s rejection of, or reclaim of, such refunds which is the same factual scenario underlying the claims in relation to the GTA appeal and the €127m appeal.

…

…I note that the trading underlying both the GTA claim and the €127m claim was OTC; and that, based on the GTA’s current practices, it is also a key factual issue with regard to the €127m appeal whether MFGUK’s former clients were (at least indirectly) acquiring from short sellers. Both sets of trades were carried out over the dividend distribution date, and so raise the same questions around beneficial ownership.”

The Administrators submit in reply, that as so explained and clarified, the close connection is not sufficiently established, for the following reasons:

In my view, it is plain that, if any claims are eventually made, it is likely that such claims will raise many issues of German law which are the same or very closely similar to those that the German Fiscal Courts are to deal with in the context of the Later MFGUK Refund Claims (and pursuant to the stay I have agreed to grant) the GTA Claim.

The most obvious example is the issue relating to the effect and application of section 44 para 1 sent. 3, para 5 ITA (see paragraphs [131] to [133] above) which was considered by the Fiscal Court of Hesse, whose decision has to some extent been the catalyst for DB’s claim. There is a difference in view between DB (and the GTA) on the one hand, and the Administrators on the other, as to whether that provision does apply in the various contexts the subject of the various claims; but the fact remains that in rejecting the DB €127m Proof/Appeal the Administrators expressly joined issue on the point, and it is undoubtedly in play: see paragraph [133(2)] above. Mr Moss’s submissions, perhaps for forensic reasons, never quite grappled with this overlap: indeed, he rather tended generally to deny it.

Further, even issues which may at first blush appear severable and dispositive, and capable of discrete adjudication accordingly, may not be so in the context. Thus, Mr Fisher submitted in reply that even on an issue which might at first blush, at least, appear to invite straightforward application of readily understandable provisions, whether of German law or English law, that is, the issue of claim limitation or time bar, may itself run into tricky questions as to the effect of the Hesse judgment on 10 March 2017 and whether the commencement of any relevant limitation period would be postponed until that date: according to paragraph 32 of Berghes 2, under German law the position differs from English law (in the context of which knowledge of the law is not necessary for the commencement of a limitation period) and “in the case of complex legal matters unresolved by German case law, the beginning of the limitation period is postponed until a final and binding court jurisprudence regarding the legal issue has been established”. Again, Mr Moss tended to emphasise the advantage of such a process of resolution without explaining the means.

However, there are other considerations to be weighed in considering whether it is sufficient, and in particular, how great is the risk of inconsistent judgments in this particular context.

First, DB is not a party, nor is it proposed to be a party, in either the GTA Claim or the Later MFGUK Refund Claim. Any ruling in those claims would have the force of precedent (at least in the mindset of this jurisdiction): but it would not bind DB. Further, the focus of any trial of the DB Claim for an indemnity would, as best one can tell given that no claim against DB which may give rise to such claim for an indemnity has yet been made, be rather different than the focus of trials of the GTA Claim and the Later MFGUK Refund Claim.

Secondly, and more generally, it is, to my mind, counter-intuitive to grant to an applicant against whom no proceedings have been brought, a stay pending proceedings to which the applicant is not a party and which will range over rather different issues as well as any identified common issues, except on the clearest demonstration that the common issues will in effect determine the result in the proceedings sought to be stayed. I do not think that would be the case, although until the shape of any ‘incoming’ claim against DB is more defined, it is not easy to be definite about this.

Thirdly, for the reasons I have sought to outline in paragraphs [62] to [69] above, there is a greater reluctance to grant a stay of an insolvency process instigated by the applicant himself and which would depart from the statutory scheme of adjudication. I must also take into account Mr Moss’s submission that, in any event, the assessment and valuation of contingent (in this case hypothetical) claims is a matter of English law and for the Administrators to determine (subject, of course, to review by the

Fourthly, my decision to refuse a stay in respect of the DB Mirror Proof/Appeal may be a factor in favour of not intervening in the context of the DB €127m Proof/Appeal either.

Fifthly, and of importance, there is no possibility of the DB €127m Proof/Appeal being determined ultimately except by this Court: for its valuation is a matter governed by the Rules to which I have referred previously, and the adjudication of this court’s officers, subject to this Court’s supervision.

To my mind, these contrasting arguments demonstrate that the decisive factors in determining whether to grant a stay (at least at this stage) are (a) whether there is any satisfactory and available means of determining the DB Proof/Appeal without having for that purpose to decide systemic, complex and hotly contested issues of German tax law, and running the real risk of inconsistent adjudications of them (which there is not in the context of the GTA Proof/Appeal, but may well be in the context of the DB Mirror Proof/Appeal); and (b) whether, in the particular circumstances of this administration, the interests of the body of creditors and of the conduct of the administration in accordance with the Insolvency Rules are outweighed by any unavoidable need to consider issues of German tax law more easily and satisfactorily determined in Germany. I turn, therefore, to address the more practical considerations posed by question (2) in paragraph [190] above and to assess how, in all the circumstances, the interests of the general body of creditors and of the expeditious and efficient conduct and conclusion of MFGUK’s administration would best be served.

In determining the interests of the general body of creditors I attach weight to the considered assessment of the Administrators. The Administrators contend that the disposal early of the DB €127m Proof/Appeal would “in and of itself, enable a further meaningful distribution to be made from the estate” and would save costs. Mr Moss submitted forcefully that DB “has no right to hold up the distribution of assets to undisputed creditors on the basis of a hypothetical future claim against it and its hypothetical future indemnity for that claim.”

Mr Moss also floated in submission (though Mr Smith protested there was no evidence to this effect (Footnote: ¹⁰)) the additional possibility that early determination could pave the way to a new proposal for a CVA, which might be especially welcome to creditors who overwhelmingly supported the previous proposal which but for the late emergence of the DB €127m Proof/Appeal would have become unconditional and resulted in the vast majority of them obtaining a further and final dividend.

Obviously also, the longer the administration of MFGUK continues, the greater the cost and the lower the final distribution. There was a disagreement between the parties

More generally, Mr Fisher’s principal point on this aspect of the matter was that the estate will have to remain open until the MFGUK tax reclaims in Germany have been resolved: it would appear (subject to clarification from the Administrators) that the running costs of £1.8m a year will be incurred whether or not the DB Stay Application is granted. He also made the point that in the failed CVA, it had been envisaged that the estate would, if the CVA did not proceed, have to be kept open for some six to eight years to enable the Later MFGUK Refund Claim to be adjudicated, and possibly 10 to 12 years if further reclaims under EU law were also pursued. To this he added that:

“the court ought not be overly concerned by any further delay in circumstances where creditors have already received 90p in the pound, all client money entitlements have been dealt with, all client assets have been dealt with.”

The comparative costs of pursuing and adjudicating the DB €127m Proof/Appeal here, rather than in Germany and then here, are not easy to assess, unless of course the matter can be concluded in some way which strips out the issues relating to the German Tax legislation: for example (and as was suggested in Pink 7 at paragraph 4.4) by determining the issues relating to the DB €127m Proof/Appeal on the assumption that the relevant claim(s) against DB for which indemnification is sought by DB from MFGUK are valid. But DB expressed concern about this, on the usual basis that seeking to determine the matter, in effect as a preliminary issue, in such a way might prove to be a short cut to a long journey. The possibility was ventilated in skeleton arguments and in paragraph 4.4 of Pink 7; but it was not substantially elaborated by Mr Moss, and Mr Fisher was dismissive of it as “unviable”. But it is, to my mind, an important consideration

Indeed, as it seems to me, in circumstances where there is presently no rival process in Germany impleading DB, nor any actual claim against it anywhere, the balance comes down to determining whether there may now be fashioned potentially dispositive issues well capable of being adjudicated by this Court without seeking to pre-determine the matters in contest in the GTA Claim and the Later MFGUK Refund Claim.

As it seems to me, the DB Proof/Appeal plainly does raise issues which do not arise in the context of the GTA Claim or the Later MFGUK Refund Claim. In particular, although common issues arise as regards whether or not DB is exposed to claims, questions relating to its entitlement to lay off or indemnify itself against such claims raise issues of law, some English, some German, which may not necessarily be enmeshed in the systemic issues of German tax law which the GTA and DB focus upon. At least some of the Administrators’ reasons for rejecting the DB Proof, including their position that in the circumstances they are entitled to ascribe a nil value, even though disputed, may be capable of preliminary determination.

I recognise that DB are resistant to such a process; and I have already referred to some of their reasons for doubting its practicality (see, for example, paragraph [198] above as to the issue of limitations). Mr Fisher submitted that the possibility of splitting the issue of any liability on the part of DB to account for WHT (DB’s exposure) from any entitlement on the part of DB to an indemnity in respect of such exposure was, in effect, (my word) a chimera, primarily because such entitlement depended on the nature of the exposure. He submitted, more particularly, that both were inter-related matters of German law, and especially the interpretation and interaction of section 44 of the German ITA and sections 670 and 675 of the BGB and their application in the context of Cum/Ex transactions (and see paragraphs [132(1)] and [133(1)] above).

Mr Fisher similarly, in addition to his submissions as to the unreality of Mr Moss’s notion of safely isolating an issue of limitations, also rightly reminded me of the usual judicial antipathy to preliminary issues as being treacherous short cuts and always difficult to formulate in a manner which ultimately secured its objective.

I bear these considerations carefully in mind. I bear in mind also the risk that the costs of a preliminary process which might not ultimately be decisive, and might then result in delay and additional expense and even (for example, in the event of further appeal) parallel proceedings. Nevertheless, and even discounting Mr Moss’s more extravagant claims as to the likely savings of time and cost, I consider that the possibility of early resolution paving the way to a further distribution of some £30 million, translating into a further dividend of 3.3p in the £ (Footnote: ¹¹), and/or to the sort of solution proposed under the CVA which fell away, should not be dismissed, and the alternative of almost inevitable long delay and increased cost is to be avoided if at all possible.

There being, as yet, no parallel proceedings impleading DB little is lost by affording time to explore the possibility of culling from the Administrators’ objections to the DB Proof issues in the nature of preliminary points to be determined on agreed or prescribed assumptions. That is the general thrust of the possibility put forward in paragraph 4.4 of Pink 7, though it has not really been carried through or elaborated yet, as now I consider it should be.

To facilitate this, and balancing all the factors to which I have referred as best I can in what remains a somewhat opaque situation where the exact basis of and issues raised by any claim against DB cannot yet be clear since none has been made, I have concluded that I should, at this stage at any rate, refuse DB’s application for a stay of the DB €127m Proof/Appeal. Refusing a stay now also has the benefit that the matter

In short, for the present I do not think that there is presently sufficient reason to displace or delay the process of proof and appeal to which DB submitted, albeit in circumstances where they had little choice. By careful selection of potentially dispositive issues, I consider that there is some prospect of that process enabling a determination without recourse to the intricacies of German tax law which are to be decided in the context of the GTA Claim; whereas an immediate stay guarantees a long delay before this court can determine the matter, based on presently hypothetical claims, after a long wait for non-binding guidance from the German court which may result from other cases to which DB is not a party.

In the later context, I would invite careful and constructive consideration as to how that may be achieved. I shall for that purpose give permission to apply. If the quest proves unworkable, then the application for a stay may be restored. My decision now should not in any way be taken as precluding the grant of a stay hereafter if it becomes apparent that the DB €127m Proof/Appeal cannot fairly be adjudicated in the manner that I have suggested may be possible.

I should presently envisage that the DB Mirror Proof/Appeal and the DB €127m Proof/Appeal may be heard together. But that and the other matters canvassed above can be further discussed after the parties have had time to digest this judgment after its circulation in draft.