This judgment is in the following parts:

Introduction

The case concerns an appeal brought by The Financial Conduct Authority (“the Authority”) against a decision of the Upper Tribunal Tax and Chancery Chamber (“the Upper Tribunal”) by Judge Timothy Herrington that the Authority pay a proportion of the costs of the two respondents, Mr Seiler and Mrs Whitestone. The Costs Decision was dated 9 November 2023, and has the neutral citation [2023] UKUT 00270. Mr Seiler and Mrs Whitestone had been parties in proceedings before the Upper Tribunal, having referred to the Upper Tribunal the Decision Notices that had been issued by the Authority against each of them. Following the Costs Decision, in which the Authority was ordered to pay some of the costs of Mr Seiler and Mrs Whitestone, the Authority sought permission to appeal on three grounds, numbered Grounds One, Two and Three. The Authority was granted permission to appeal by Falk LJ on Grounds One and Three only, and therefore renumbered those Grounds One and Two. Ordinarily, it would not be necessary, on an appeal, to refer to the substance of a ground for which an appellant did not obtain permission. However, in this case the substance of that original ground which was refused permission is also of relevance, and I will refer to that as the Refused Ground below.

The Authority had, through its Regulatory Decisions Committee (“RDC”) issued Decision Notices to each of three individuals, namely Mr Seiler, Mrs Whitestone and another person, Mr Raitzin. Those Decision Notices, had they led to Final Notices, would have prohibited each of those three individuals from performing any function in relation to any regulated activities carried on by an authorised or exempt person, or exempt professional firm, pursuant to section 56 of the Financial Services and Markets Act 2000 (“FSMA 2000”). The basis of each of the Decision Notices was that the Authority had concluded that each of the three had acted recklessly and with a lack of integrity in respect of certain events that had arisen from the relationship between their employers, which were all members of the Julius Baer banking group of companies (“Julius Baer”), and someone called Mr Merinson, who was connected with the Yukos group of companies (“Yukos”). Mr Merinson had received a large proportion of commission paid in respect of different foreign exchange transactions. Each of the three executives referred their respective Decision Notices to the Upper Tribunal.

Mrs Louise Whitestone was employed as a relationship manager by the UK subsidiary of Julius Baer, on its Russian and Eastern European Desk from January 2009 until November 2012. Mr Thomas Seiler joined Julius Baer in 2008 as a managing director, and was part of its senior management team. He became the Sub-Regional Head for Russia and Eastern Europe, a role that he occupied until 2014. The third applicant before the Upper Tribunal in the main proceedings was Mr Raitzin. He was a member of the bank’s executive board, and its Regional Head for Latin America.

The Upper Tribunal conducted a hearing which lasted 20 days and ran between November 2022 until January 2023. In a judgment at [2023] UKUT 00133 (TCC) dated 12 June 2023 the Upper Tribunal, chaired by Judge Timothy Herrington, found that the Authority had not made out its case that each of the three had acted recklessly and consequently with a lack of integrity in relation to the subject matter of the references. The issue of whether the prohibition contained in the Decision Notices should be imposed upon any of the three was remitted by the Upper Tribunal to the Authority for it to reconsider its decisions. Shortly after that, the Authority chose to discontinue the proceedings and withdraw the Decision Notices.

I observe as a preliminary point that the substantive decision by the Upper Tribunal referred to at [5] above runs to over 1,000 paragraphs and is, as one would expect given the length of the hearing, extraordinarily detailed. There are 27 different named individuals in the decision, a glossary and an annex explaining company structure. It is not necessary to provide more than an outline summary of it, but this subject is something to which I will return.

In granting permission to appeal on the two grounds, the single Lady Justice imposed a condition that the Authority should bear its own costs of the appeal in any event. Correspondence with the court had indicated that neither of the two respondents, Mr Seiler or Ms Whitestone, were in a financial position to pay any costs for representation such that they could appear on any appeal. Mrs Whitestone in particular had not had the benefit of Director and Officers’ Insurance and had incurred very considerable personal legal costs in the Upper Tribunal. Attention was drawn by the single Lady Justice in her order granting permission, to the corresponding duty upon the Authority to draw the court’s attention to matters that would undermine its case. Ms Clarke KC and her instructing solicitors appeared for her before us pro bono, having come to the conclusion that the Authority’s skeleton argument did not sufficiently address points adverse to it. This court therefore had the benefit of submissions on Ground One, which also concerned Mrs Whitestone as well as Mr Seiler. We did not have the benefit of submissions from the respondent on Ground Two, as Mr Seiler did not appear and was not represented on the appeal.

Following the hearing itself, on 17 May 2024 the Authority submitted a further decision of the Upper Tribunal which had been handed down on 14 May 2024, namely Banque Havilland SA and others v Financial Conduct Authority [2024] UKUT 01115 (TCC). This was done in order to make sure that the court had all potentially relevant authorities before it, and in particular because Mr Seiler was not represented. I address that decision further below at [64].

Legal Framework

The Financial Services Act 2012, amongst other things, changed the name of the Authority from the Financial Services Authority, the name of its predecessor. The Authority is responsible for regulation of financial firms and maintaining the integrity of financial markets in the United Kingdom; it is therefore the regulator. Although the proceedings arose out of events taking place between June 2009 and February 2012, nothing turns upon the change from the Financial Services Authority to the Authority in this case.

Part V of the Financial Services and Markets Act 2000 (“FSMA”) is the relevant statute that deals with the performance of regulated activities under the Act, and section 56 within that part deals with prohibition orders. The Decision Notices would have led to the imposition of prohibition orders upon each of the three executives to which I have already referred. Under section 56(1) the Authority may make a prohibition order:

If a person subject to a prohibition order performs a function in breach of it, they are guilty of an offence under section 56(4) FSMA and liable on summary conviction to a fine. Under section 57, if the Authority proposes to make a prohibition order it must first issue a Warning Notice, and then after considering the response of the person in question it may issue a Decision Notice, deciding to impose a prohibition order. If it does so – and the Authority for these purposes acts through its Regulatory Decision Committee – then under section 57(5) the recipient of a Decision Notice has the right to refer the matter to the Tribunal.

Section 133 FSMA governs the role of the Tribunal, which may be a disciplinary or non-disciplinary reference. Section 133(7A) lists those matters which are disciplinary references; those which are the subject of these references and of this appeal were non-disciplinary references. Under section 133(4) the Tribunal may consider any evidence relating to the subject-matter of a reference or appeal, whether or not it was available to the decision-maker at the time. It is not therefore an appeal against the Authority’s decision, but a re-hearing of the issues which gave rise to the decision.

The powers of the Tribunal are limited upon such a reference. For a non-disciplinary reference such as this one, they are set out in section 133(6) and are as follows:

This is entirely consistent with the role of the Authority as a regulator, and the purpose of the regulatory regime, which is to maintain the integrity of financial markets.

Section 29 of the Tribunals, Courts and Enforcement Act 2007 (“TCEA”), which deals with the costs regime in tribunal proceedings generally, provides that:

Section 29(3) TCEA makes it clear that the power to award costs is subject to the Tribunal Procedure Rules. The relevant rules in this case are Rules 10(3)(d) and 10(3)(e) of The Tribunal Procedure (Upper Tribunal) Rules 2008 (“the Rules”) which state the following:

There are a large number of tribunals of different types which deal with a great many different areas and kinds of subject matter. The costs regime for tribunals is different to that of court proceedings, for example. Costs are not routinely awarded in tribunals, and part of the ethos of this is the different nature of the jurisdiction of tribunals. This was a financial services case.

These provisions make it clear that whether or not costs are to be awarded in any particular case and, if so, of what amount, is a matter of discretion for the Tribunal. Like all judicial discretions, it must be exercised having taken account of all relevant circumstances and ignoring all irrelevant factors. Given the express terms of Rule 10(3)(d) and (e) (the latter only applying in a financial services case, which this one is) unreasonableness is required. This was correctly described by Ms Clarke KC as a “threshold condition” for any order of costs. That term has its origins in the case of HMRC v Jackson Grundy [2017] UKUT 180 (TCC) at [47], in which the Upper Tribunal summarised the principles to be applied when considering Rule 10(3)(d) as follows:

The consideration of discretion only comes after the relevant finding that permits that discretion to be considered and exercised. Unreasonableness is effectively a condition precedent to the jurisdiction to order costs being triggered. Absent any finding or conclusion by the Upper Tribunal of unreasonableness, there would have been no jurisdiction on the part of the Upper Tribunal to have ordered costs. In this case, when considering the application for costs, the Upper Tribunal concluded that the threshold condition was satisfied and that the Authority had acted, in some respects, unreasonably, and therefore awarded costs. The Authority was ordered to pay all of the costs of one particular issue, and 5% of the remainder of the costs of each of Mr Seiler and Mrs Whitestone.

Factual background

Bank Julius Baer & Co Ltd is a private bank incorporated and regulated in Switzerland. It has various subsidiaries across the world, and Mrs Whitestone was employed as a relationship manager by Julius Baer International Ltd, Julius Baer’s UK regulated subsidiary. Mr Seiler was employed as the Sub-Regional (Market) Head for Russia and Central and Eastern Europe at the Zurich office and was the functional line manager for Mrs Whitestone. For completeness only, Mr Raitzin was also employed in Switzerland and was Mr Seiler’s line manager. He did not seek his costs below, is not a respondent to this appeal and therefore it is not necessary to deal with him further.

The Yukos Group of companies includes Yukos Capital S.a.R.L, Yukos International UK BV, Yukos Hydrocarbons Investments Ltd and Fair Oaks Trade and Investment Ltd. Mr Dmitri Merinson was the Financial Controller of Yukos International and the Chief Financial Officer of Yukos Capital and Yukos International. Yukos was an oil and gas company of considerable size based in Russia which no longer exists, having been broken up by the Russian state.

The Authority issued the Decision Notices because of conduct in relation to arrangements entered into between Julius Baer and Mr Merinson, whereby it was anticipated that he would introduce companies within the Yukos Group to banks with the Julius Baer group and receive remuneration for doing so.

Payments were made to Mr Merinson by Julius Baer between 2010 and 2012 following the execution of three separate foreign exchange (“FX”) transactions for Yukos, referred to as the First FX Transaction, the Second FX Transaction and the Third FX Transaction. However, such transactions were not the only type of business considered within the substantive decision, which also contained analysis of other financial dealings, including what was called a “CoY”, a derivative instrument combining an FX linked deposit with a currency option, with the aim of providing a higher yield return than that available for a standard deposit, but which also carried a higher risk than a standard deposit due to the exposure to FX rate movements. The substantive decision considered a complex history of financial dealings over a considerable period of time.

The Authority considered that Julius Baer’s conduct in its relationship with the Yukos Group had demonstrated a lack of integrity, and this was why the Decision Notices were issued. The Authority’s case was that Julius Baer must have appreciated the clear risk that by entering into the arrangements that it did with Mr Merinson, on the terms carefully explained in the substantive decision, it might be facilitating or participating in financial crime.

In summary only, the Authority had concluded that the three executives had disregarded risks of which they were actually aware that Julius Baer might be facilitating or participating in financial crime by entering into these arrangements. In its decision on the references to it, the Upper Tribunal found that it would be irrational of the Authority to make a prohibition order against any of the three on the basis that they had acted without integrity. The matter was remitted to the Authority for reconsideration on the basis of the findings of fact that were set out in the substantive decision. The Upper Tribunal did find that there were a number of instances in which each of the three applicants in the case had demonstrated varying degrees of a lack of competence and capability. However, the Upper Tribunal found that a prohibition order should not be considered as a proxy for a disciplinary sanction in circumstances where the imposition of a disciplinary sanction against any of the three applicants concerned could not be imposed, either because he or she was not an approved person under FSMA 2000 at the relevant time or, where he or she was an approved person, the relevant limitation period had expired.

So far as the Third FX Transaction in particular was concerned, the Upper Tribunal took a dim view of the fact that the details of what was said to be that transaction had changed from the Warning Notice stage, and that the details of what was said to be the Third FX Transaction were different in the Decision Notices issued to each of Mr Seiler and Mrs Whitestone than they had been initially. The Upper Tribunal concluded that there were serious failings in the way that the Authority had conducted its investigation. It found that there was no good reason why the correct facts and matters regarding what the Upper Tribunal called the “New Third FX Transaction” (by which it meant, the different case on that transaction advanced by the Authority at the substantive hearing) could not have been contained in the Warning Notice, had the Authority conducted its investigation with due skill, care and diligence. It also found that the Authority had the opportunity to correct its earlier mistake by asking the RDC to issue a revised Warning Notice with the correct details later relied upon by the Authority, which would have permitted the RDC to test the allegations first by seeking the response from those concerned. The Enforcement section of the FCA had proposed that the Warning Notices be amended in this way, but that had not been done. Matters had then proceeded as if an amendment had been made concerning this transaction, but the RDC had not made that amendment and the proposal by the Enforcement section was not carried through.

The Upper Tribunal found that it was therefore procedurally irregular for the RDC to have made the findings it did without the Warning Notice having been amended in respect of the Third FX Transaction. There were a range of other findings made, but essentially for the purposes of the Costs Decision the Upper Tribunal did not accept that Mr Seiler and Mrs Whitestone had not been prejudiced by what had happened concerning the change in case regarding the Third FX Transaction. It found that they did not have the protection of the matter having properly gone through the Warning Notice procedure, when that was clearly an option open to the Authority at the time. It also found that “It is clearly not appropriate for new allegations to be introduced during the course of the representations phase after the investigation should have been and could have completed” (at [1020] of the substantive decision). The precise terms of the transaction which was alleged to evidence a lack of integrity should have been properly formulated.

Both Mr Seiler and Mrs Whitestone made applications for costs against the Authority, on the basis that it had been unreasonable for the Authority to have issued the Decision Notices at all. As explained at [17] above, a finding of unreasonableness is a threshold condition to the Upper Tribunal being entitled to award costs.

The Upper Tribunal found in the Costs Decision that the FCA had not acted unreasonably in making the decisions referred or generally, but that it had acted unreasonably in certain specific respects in defending the references in the Upper Tribunal. It had been necessary to consider the evidence of Mr Seiler and Mrs Whitestone in order to arrive at the conclusion that they had been lacking in competence, rather than that they had lacked integrity (the latter being the relevant test under the regulatory regime). Indeed, Mr Seiler had conceded in his evidence before the Upper Tribunal that he had been incompetent, and Mrs Whitestone had accepted that she had been both naïve and incompetent.

However, Upper Tribunal Judge Herrington did find at [146] of the Costs Decision that the Authority had been unreasonable in certain respects and that he therefore had jurisdiction to make a costs order in respect of both the Third FX Transaction, and also the other matters that were listed at [116] of the Costs Decision. These other matters were that the Authority had unreasonably failed to call material witnesses; that the Authority had refused to answer Mr Seiler’s requests for clarification about what other individuals were said to have known about the transactions in question at the time, and the relationship between Mr Merinson and another employee at Julius Baer; that the Authority had repeatedly declined to respond to Mr Seiler’s repeated requests for information and documents concerning its investigation, and the evidence called by the Authority to explain its document gathering processes exposed significant gaps and flaws; and the Authority’s decision to continue to rely in the proceedings before the Upper Tribunal upon the Third FX Transaction.

Given this conclusion of unreasonable conduct by the Authority, the Costs Decision recorded the conclusion of Judge Herrington on this issue and stated the following conclusions:

The Judge therefore made costs orders against the Authority in the amounts which he then came on to consider at [152] and following under the heading “Quantum”. That point does not arise on this appeal, and the challenges that are brought by the Authority are in respect of his findings regarding unreasonableness. There is no challenge to his assessment that the correct award was 5%, nor (given the refusal of permission to appeal on the Refused Ground) to his conclusion regarding the Third FX Transaction.

The Grounds of Appeal

These were renumbered following the refusal of permission for one of them. Those before us are as follows. The paragraph references to each were included by the Authority in its grounds of appeal:

In view of the materiality of those paragraphs of the Costs Decision to each of the two grounds, I shall reproduce them in full.

The relevant passages of the Costs Decision specifically identified in the Grounds are as follows:

The Challenges to the Findings of Unreasonableness

Whether a party, either the Authority in this reference or indeed anyone, has acted unreasonably, can only sensibly be determined after consideration of all the factual circumstances. This is an evaluative assessment that can only be reached after detailed analysis of the facts. Upon appellate review in any type of case, the court would generally not interfere with the conclusion of the court or tribunal below on findings of fact.

However and in any event, section 13(1) of TCEA states the following:

In this case, the Costs Decision is not an excluded decision and section 13(14) does not apply. This therefore means that any appeal can only be considered under section 13(1) TCEA. The Authority is only permitted to appeal the Costs Decision on a point of law, and this is expressly stated in that section. The scope of section 13(1) has previously been considered by this court. Obrey and others v Secretary of State for Work and Pensions [2013] EWCA Civ 1584, [2014] HLR 12 concerned an appeal from the Upper Tribunal Administrative Appeals Chamber. The subject matter was housing benefit for accommodation and the withdrawal of that benefit from persons absent from their home for in excess of 52 weeks (because they could not be treated as occupying the accommodation as their home). A challenge was brought by mentally ill patients detained under the Mental Health Act. The First-Tier Tribunal upheld that challenge; the Secretary of State appealed and the Upper Tribunal over-turned the decision, and re-made it, finding indirect discrimination but that it was justified.

On appeal by the patients to this court, at [13] to [18] Sullivan LJ considered the scope of section 13(1) TCEA and explained the following:

Before us, the Authority challenged the approach of the Upper Tribunal in two ways, and it also cited Various Eateries Trading Ltd v Allianz [2024] EWCA Civ 10, in particular [57] to [59]. Its primary route was that the Costs Decision disclosed some error of principle. In the alternative, it maintained that the basis of the Costs Decision, in its assessment of reasonableness, was not one that was reasonably open to the Upper Tribunal.

The first important point to note is that the Various Eateries case concerned an appeal to this court (Sir Julian Flaux C, Newey and Males LJJ) from a hearing of preliminary issues by Butcher J in the Commercial Court. These were issues relating to the coverage of a business interruption policy on the Marsh Resilience form, and the degree to which this policy provided coverage to the closure of a chain of restaurants during the Covid-19 policy. Butcher J heard expert evidence and had made certain factual findings, set out at [87] in his first-instance judgment at [2022] EWHC 2549 (Comm) and summarised at [8] by Males LJ. After considering the causation question that the judge had undertaken, which required an exercise of judgment in particular in respect of what is called “original cause” in the insurance field and remoteness, Males LJ stated the following:

These passages concern challenges to evaluations of fact and in my judgment do not assist the Authority. Firstly, they do not concern an appeal from the Upper Tribunal at all. Expressly and by statute, there is no challenge permitted on appeal to any findings of fact by an Upper Tribunal. Appeals are only permitted on points of law, by reason of section 13(1) TCEA. Secondly, they concern appeals on fact from a first instance High Court judgment, and not a costs decision, and the case is therefore plainly distinguishable on that basis. Thirdly, they emphasise that even on that test – which I do not consider to be the correct one, given the restriction in section 13(1) to points of laws only – there would be a very high hurdle for the Authority to overcome on appeal. Finally, it might equally be said, by analogy with the dicta at [15] in Obrey, that the particular expertise of the Upper Tribunal concerning financial regulation and both disciplinary and non-disciplinary references ought to be recognised, and it ought – in principle - to be even more difficult for evaluative fact to be challenged on appeal in any event, even if that were permissible. However, it is not permissible, due to the restriction on appeals of fact that arises under section 13(1) TCEA.

Further, given the nature of the threshold condition, which must by its nature inevitably consider and take into account a great amount of fine factual detail of the proceedings that any appellate court would find difficulty in matching, this means in my judgment that the “pragmatic approach” referred to by Lord Hope in Jones must lean heavily towards the expertise of the Upper Tribunal in these matters in any event.

The Authority, in its skeleton argument, submitted that the Costs Decision adopted what was described as “a novel and highly unorthodox approach in respect of important matters of practice and procedure in the Upper Tribunal. Those issues include, inter alia, the extent to which the Authority must call witnesses that it cannot proffer as witnesses of truth and whether it must invite the Tribunal to use potential powers to summon witnesses from abroad – the use of which is hitherto unprecedented – to avoid a finding of unreasonableness for costs purposes. The Costs Decision also appears to require the Authority to provide information about the subjective states of mind of individuals which it does not possess. The Authority disputes the correctness of the Tribunal’s findings, which are wrong in law; additionally or alternatively, in each case, contingent on a false premise as to the proper role of the Tribunal, and the Authority, as the regulatory body before it, in adversarial proceedings.” The Authority also stated in its written skeleton concerning the findings against it, that “the criticisms of the Authority upon which the Costs Decision is premised…..are unfair and unprincipled.”

In oral submissions, Mr George KC for the Authority sought to amplify what he characterised as a failure of understanding on the part of the Tribunal in respect of the very nature of proceedings before it, when a party made subject to a Decision Notice sought to challenge it by referring it to the Upper Tribunal. On several occasions, he submitted that the Authority should be treated as though it were a litigant in ordinary civil proceedings in the High Court, and that how it conducted itself, both in the investigation and reference before the Upper Tribunal, should be considered against that standard. He also relied upon some isolated references within the Costs Decision itself where an analogy had been drawn by the Upper Tribunal Judge between the nature of proceedings before it, and those in the Administrative Court of the High Court, where the Costs Decision had described its jurisdiction as being “akin to judicial review”. A range of analogies were drawn by Mr George between the approach of the Administrative Court when considering such an application, with how the Upper Tribunal ought to conduct itself, in an attempt to demonstrate that the Upper Tribunal had misunderstood the proceedings and therefore erred in reaching the conclusions in the Costs Decision, which amounted to errors of law.

It is common ground between the parties to this appeal that, in relation to non-disciplinary references, the Upper Tribunal exercises a supervisory jurisdiction only. This means that, although it must reach a fresh decision as to the facts and there is no limitation on the evidence that may be put before it (it is not restricted by what the decision-maker knew at the time), it cannot substitute its own decision for that of the Authority. All it can do is remit the matter for the Authority to reconsider its decision in view of the Tribunal’s findings. This is well-established and set out, inter alia, in Financial Solutions (Euro) Ltd v Financial Conduct Authority [2020] UKUT 243 (TCC) (“FSE”) at [21], itself citing the earlier case of Carrimjee v Financial Conduct Authority [2016] UKUT 0447 (TCC).

On the basis that the Upper Tribunal has a supervisory function, and cannot substitute its decision for that of the regulator, with this limited power of remission, I can understand the reason for the terminology adopted in the Costs Decisions and some of the similarities drawn with the role of the Administrative Court in judicial review. I shall return to the use of that analogy at [75] and [76] below. However, rather than concentrate on that analogy, there are other decisions, both of the Upper Tribunal, where the particular role and approach of the Upper Tribunal has been considered.

In Frensham v Financial Conduct Authority [2021] UKUT 0222 (TCC) Mr Frensham had referred the Decision Notice issued against him by the Authority to the Upper Tribunal. He had been convicted in 2016 of sexual grooming of a child under the age of 16, and given a suspended sentence of imprisonment. The Authority concluded that he was not a fit and proper person. The Warning Notice was issued in 2020, evidently some considerable time after his conviction. One of the issues that came to light in the evidence before the Upper Tribunal were discussions on policy within the Authority (at a level higher than that of the Authority’s witnesses who had been called, one Mr Blunt and Ms Couzens) which were of potential relevance to this period of delay. In respect of this, the Upper Tribunal said as follows:

In Financial Solutions (Euro) Ltd v Financial Conduct Authority [2020] UKUT 0243 (TCC) the Upper Tribunal considered an application for costs following a reference to it by Financial Solutions (Euro) Ltd in which the Upper Tribunal had allowed that reference and remitted the matter for reconsideration. Having considered the necessity to find unreasonableness in order to permit an award of costs, the decision of Tribunal Judge Herrington at [29] included the following statement: “I also consider that another example of where a party could be held to have acted unreasonably is where it fails to produce evidence to support its case”.

In Forsyth v (1) Financial Conduct Authority (2) Prudential Regulation Authority [2021] UKUT 0162 (TCC), Mr Forsyth had been made subject to two prohibition orders and had a financial penalty imposed upon him. He referred these matters to the Upper Tribunal. At [65] and [76] the Upper Tribunal criticised the absence of witnesses who were not called by the regulators, who might have been able to assist the Tribunal, and drew adverse inferences from their absence.

Essentially, the submissions of Mr George KC in this respect are that it is entirely a matter for the Authority which witnesses it calls, and that it amounts to an irrelevant consideration when assessing unreasonableness whether the Authority has called certain individuals or not. I am unpersuaded by those submissions. In my judgment, the Authority is engaged in a common enterprise with the Upper Tribunal in ensuring that the objects of the legislation are achieved and that public confidence is maintained in the integrity of financial markets, with those who are not fit and proper persons prohibited from engaging in regulated activity. Support for this conclusion is found, for example, in section 133A(5) FSMA 2000 which permits the Upper Tribunal, on determining a reference to it in respect of a decision by the Authority, to “make recommendations as to its regulating provisions or its procedures”. Those recommendations can only be to the way in which the Authority will act in the future, including on other cases; this shows that the proceedings before the Upper Tribunal are not precisely the same as “ordinary civil litigation”.

However, regardless of that (and the validity or otherwise of the analogy with judicial review), I do not consider that the conclusions reached by the Upper Tribunal Judge on unreasonableness of conduct by the Authority can accurately be characterised as errors of law.

Returning to the pragmatic approach that should be taken to the dividing line between law and fact, it is clear to me that when considering potential unreasonableness both as to the conduct of an investigation and/or the proceedings before the Upper Tribunal, it is that latter body that is ideally placed to come to a considered conclusion on this threshold condition. The Authority is clearly very much of the opposite view; the skeleton argument describes the conclusions in this case as “unfair and unprincipled”. Not only do I disagree with that description, but to describe it as involving an error of law is to push the dividing line clearly beyond what could be said to be a point of law and firmly into the province of evaluative fact.

In my judgment, the challenges contained in the Grounds of Appeal must be considered and assessed by detailed analysis of the full Costs Decision. It is said in Ground One that the Upper Tribunal erred in finding that the Authority acted unreasonably in failing to call material witnesses. However, that is not what the Costs Decision stated, and is not the conclusion that underpinned this finding.

Certain witnesses were overseas in Switzerland, namely Ms Bielmann and Mr Courrier, and other important witnesses were not called, most notably Mr Campeanu and Mr Narrandes. There was a difficulty with Mr Campeanu in that the Authority had concluded, in advance of the hearing, that he could not be tendered as a witness of truth. But to portray the findings in the Costs Decision as being that “the Authority must call witnesses that it cannot proffer as witnesses of truth and whether it must invite the Tribunal to use potential powers to summon witnesses from abroad” puts the matter on a wholly different footing than the one adopted by the Judge in his findings in the Costs Decision.

What the Costs Decision stated in this respect can be summarised as follows:

It was these matters that led to the conclusion at [124] of the Costs Decision that “the Authority took no steps to demonstrate to the Tribunal that it had tried to obtain all relevant evidence.” Mr George resisted very strongly any suggestion that the Authority stands in any different role to what he called a party in ordinary litigation. The general tenor of his submissions was that it was entirely a matter for the Authority to decide how to put its case, which witnesses to call, and how it could and should (turning to the next ground) answer requests for information or clarification. He also submitted that because some of the witnesses were in Switzerland, outside the jurisdiction and unable to assist, there was nothing that the Authority could do.

I am not persuaded by those submissions. The Authority has a regulatory function, whose purpose is to maintain integrity and confidence in financial markets. One of the consequences of a Prohibition Order is that a person subject to such an order may commit a criminal offence if they engage in behaviour which would, absent the order, be entirely legal, but by virtue of the order they are prohibited, under risk of a sanction (a fine), from performing. The Authority is not an “ordinary litigant” and the Upper Tribunal is not conducting “ordinary civil proceedings in the High Court”.

The Court of Appeal has previously considered the issue in Financial Conduct Authority v Hobbs [2013] EWCA Civ 918. These proceedings concerned a prohibition order against Mr Hobbs, who had traded in coffee futures, on the basis that he had engaged in market abuse. One of the issues was whether Mr Hobbs had lied both to his employer and the Authority during the course of the investigation into his conduct.

In the course of determining the appeal from the findings of the Upper Tribunal (that had been made in Mr Hobbs’ favour), at [38] Sir Stanley Burnton stated the following:

In R (Wilford) v FSA [2013] EWCA Civ 677 (the FSA being the Financial Services Authority, the precursor to the FCA as regulator before the change in 2012 when the Financial Services Act 2012 came into force) the High Court (per Silber J) had quashed a Decision Notice against Mr Wilford, the Group Finance Director of the Bradford & Bingley Bank. In the context of considering Mr Wilford’s rights to bring judicial review, rather than proceeding by way of reference to the Upper Tribunal, Moore-Bick LJ stated:

These cases both make it clear that the Upper Tribunal is an expert tribunal exercising a judicial function; and that it is also an integral part of the regulatory scheme established under the Act.

To accede to Mr George’s submissions that the FCA is free to act in the way criticised by the Upper Tribunal in this case, and that such proceedings are simply “normal litigation” would be to take a different view than that contained in both the dicta of Sir Stanley Burnton in Hobbs, and also that of Moore-Bick LJ in Wilford. Even if this court were permitted to do so -which it is not, due to the doctrine of stare decisis – such a divergence from those conclusions would not be justified. Proceedings before the Upper Tribunal such as these are not ordinary civil litigation. Ordinary civil litigation between two parties, individuals or companies, does not have the added public interest in “….ensuring, so far as possible, that persons who are not fit and proper persons to perform functions in relation to a regulated activity are precluded from doing so”, nor that persons who are not so categorised are permitted to engage in such activity. In my judgment, this is an important distinction.

In the decision provided to this court on 17 May 2024, namely Banque Havilland SA and others v Financial Conduct Authority [2024] UKUT 00115 (TCC), handed down after the hearing before us had taken place on 1 May 2024, Upper Tribunal Judge Timothy Herrington considered references to the Upper Tribunal by parties, including the appellant bank, in respect of Decision Notices issued to it and other parties. The subject matter of the Decision Notices was advice given by the bank in a presentation which the FCA considered recommended manipulating trading strategies, which had it occurred in the UK would have been a criminal offence. The FCA concluded that the bank and other individuals had acted without integrity in relation to the presentation.

At [75] the judge cited the above passage from Wilford, and also another one at [37] where in the context of further explanation Moore-Bick LJ continued:

Although in the context of considering whether judicial review was, in a routine case, even available, the nature and scope of the Upper Tribunal proceedings was followed. These were specifically not ordinary litigation, as this dicta makes clear. At the case management stage, one of the matters in Banque Havilland considered by the Upper Tribunal was whether it should itself issue a witness summons of its own motion requiring a Mr Weller to give evidence as a “neutral” witness; and also whether the tribunal should also issue a witness summons or make an order or direction under Rules 16 or 6 of the Rules requiring the lead investigator to be tendered to the bank and the other applications for cross-examination. Although on the facts that application failed, it was the FCA who had invited the Tribunal to call Mr Weller as set out in [92].

There are two other important passages that relate to this ground contained in the substantive decision itself. The first is as follows:

The second is as follows:

The Authority submitted that the findings by the Upper Tribunal in the Costs Decision in respect of witnesses were wrong in law, or were “contingent on a false premise”, namely that “the Authority is required to exhaust all options to obtain evidence from any person that might reasonably be considered able to give relevant evidence to the Tribunal”. That is not an accurate summary, in my judgment, of the findings, when they are read in their proper context in the Costs Decision as a whole. Dealing solely with Mr Campeanu, the Upper Tribunal was critical of the reliance that the Authority had placed upon an email from him, had based its case theory upon that, and then continued to do this notwithstanding that it later realised it could not rely upon him as a witness of truth. That cannot be said to be wrong in law, or to be based upon a false premise. It is a factual conclusion on how the FCA had chosen to conduct itself.

In its skeleton argument in this appeal, the Authority stated that Mr Campeanu “could not generally be put forward as a witness of truth”. That may well have been – and certainly the Costs Decision found that it was – a “tactical decision” because to have done so would undermine the Authority’s case. Reliance is placed by the Authority upon two 19^th century cases dealing with impeachment of one’s own witnesses absent an application to treat them as hostile. However, such an approach is somewhat wide of the mark, in my judgment. Given the centrality to the investigation of what Mr Campeanu had initially said to the Authority, and the relevance of that to the references as a whole, the Judge found the decision not to bring this to the attention of the Upper Tribunal and seek directions, to be unreasonable. Rule 5 of the Rules deals with case management and gives the Upper Tribunal wide powers. This includes at Rule 5(3)(d) the specific power to “permit or require a party or another person to provide documents, information, evidence or submissions to the Upper Tribunal or a party”. Rule 16 also gives the Upper Tribunal the ability to call witnesses of its own volition.

Rule 16 states:

I find it difficult to see how the Upper Tribunal can properly consider acting (or not) on its own initiative as permitted under Rule 16(a) if, as here, the FCA takes tactical decisions of the type deprecated in the substantive decision at [924] in the emphasised passage above. Further, if such tactics are permitted, then in any particular case the Upper Tribunal’s exercise of its powers under the Rules is capable of being influenced by what it does, and does not, know, based upon what the FCA as regulator chooses to tell it. I do not consider the interpretation of the Upper Tribunal’s role contended for by the Authority on this appeal to be consistent with the purpose of the regulatory framework.

In Banque Havilland, the arguments advanced by the FCA as to why Mr Weller should be called by the Upper Tribunal under Rule 16, rather than the FCA call him, are set out at [95] in that judgment and were as follows:

This is, in my judgment, exactly the issue in respect of which the Upper Tribunal was deprived of having the opportunity to consider in respect of Mr Campeanu by reason of the way the FCA had acted (or not acted).

Finally on this point, and I identify this as supplementary support for my reasoning, rather than a foundation stone upon which the reasoning rests, the FCA criticises the Judge for stating in a number of places in the Costs Decision that the duty upon the FCA is similar to that of a duty of candour in judicial review proceedings. In Wilford, which was an appeal from the High Court quashing a decision under judicial review, Moore-Bick LJ restricts at [37] judicial review to “the most exceptional case”. In such an exceptional case, the FCA would in judicial review be under a duty of candour. It is difficult to see the nature of the duty as morphing or arising only in some cases (albeit exceptional ones), but not others.

The judge below did refer to the analogy with the duty of candour, but I read those comments as his attempt to explain that the FCA in such proceedings does not stand in the same shoes as an ordinary litigant in habitual adversarial proceedings. The position both of the FCA, and the Upper Tribunal, is aligned with the purpose of the regulatory regime. The duty may have some similarities with those upon a prosecutor in a criminal case, who owes higher duties to the principle of fair prosecution and the administration of justice. The presence of the express power in the Rules for the Tribunal to call witnesses of its own motion means that it is unnecessary to consider the circumstances in which a criminal court has the power to call witnesses itself depending upon decisions that a prosecutor takes in respect of whom to call. These principles are explained in R v Wellingborough Justices, ex p Francois (1994) 158 JP 813, and R v Haringey Justices, ex p DPP [1996] QB 35, both recently cited with approval by the Divisional Court in DPP v Barton [2024] EWHC 1350 (Admin) (per President KBD and Saini J). But there is only so far that analogies assist in any event, and in my judgment, they do not necessarily assist very far.

The Upper Tribunal Rules also, in my judgment, could be described as having parallels with those expected to support a quasi-inquisitorial function. Certainly they are consistent with the joint purpose both of the Upper Tribunal, and the FCA as regulator, to ensure that the integrity of the financial markets is protected and confidence in them is maintained; that persons who are not fit and proper persons are prohibited from carrying on business in those markets; and that persons who are fit and proper are permitted to do so. This is not, in my judgment, wholly the same as what the FCA describes as “normal litigation”.

The conclusion in the Costs Decision that the position regarding Mr Campeanu should have been brought to the attention of the Upper Tribunal, so that directions could be considered, and the failure to do so was unreasonable, cannot be equated to a point of law, or to have been based on a false premise. It is a conclusion based on the facts of this case. Similar observations can be made in respect of the other matters concerning the failure to call the other witnesses. The Tribunal was well aware that two of the witnesses were in Switzerland. However, these witnesses were only asked to assist fairly late in the process, and the letter from the Authority asking for their assistance did not even make an unequivocal offer to pay their reasonable travel expenses, were they to travel to London. All that was said to them regarding expenses was that reimbursement would be considered. The Tribunal Judge found these efforts to be “feeble”. A number of similar terms could be adopted.

However, in my judgment, these findings are part of the factual landscape that the Judge was entitled to take into account, and did take into account, in reaching his ultimate conclusion as to unreasonableness in the Costs Decision. When one considers the role of the Authority, the nature of the Upper Tribunal’s function when a matter is referred to it, the effect of a Prohibition Order upon an individual and the purpose of the legislation which is to maintain public confidence in financial markets, these findings are justified. But in any event, they are findings within the range of findings properly open to the Upper Tribunal. There is no error of law contained in them.

The conclusions regarding witnesses reached by the Judge in the Costs Decision are such that they are eminently within the range of decisions properly open to him on the subject of unreasonable conduct by the Authority. In my judgment they are factual. Despite the efforts of the Authority to portray them as having been reached in the context of a legal misunderstanding by the Judge of the nature and function of the proceedings before the Upper Tribunal, there was no such misunderstanding, nor was there any discernible error of law. I would dismiss Ground One.

Ground Two is that the Upper Tribunal erred in finding that the Authority had acted unreasonably by (1) refusing to respond to Mr Seiler’s request for clarification in respect of the Authority’s position on what other individuals knew about the various FX transactions giving rise to the proceedings, and the relationship between Mr Merinson and Mr Feldman; and (2) refusing to answer certain other requests for details of its investigation.

This Ground only concerned Mr Seiler. He had, through his solicitors, made requests for clarification, some months prior to the hearing of the references, about what position the Authority took regarding other individuals at the time, what they may have known about the transactions at the time, and the relationship between Mr Merinson and Mr Feldman. There were a number of these, as the initial refusals by the Authority were met with further requests, but one example of the requests and response will suffice for present purposes, and it is as follows:

The request was:

The answer was as follows:

There were a number of other requests in similar form and they all received either the same, or very similar, non-responses. The Authority therefore simply refused to address these requests. Yet, as one of its own witnesses at the hearing, Mr Neary, accepted and the Tribunal found in its substantive decision at [153] “in assessing whether a person’s reaction showed a lack of integrity, it is relevant to look at how other people reacted at the time”. There is no error of law contained in that conclusion, which is entirely sensible and correct.

In the Costs Decision itself, at [129], the Judge stated “I find the Authority’s response to these requests to be unreasonable. I accept Mr Seiler’s submissions that the Tribunal would have been assisted had the Authority given a more positive response, because, as Mr Neary [a witness for the Authority] accepted in his evidence, it is relevant to look at how other people reacted at the time.”

Requests for further information or clarification may be made for a wide variety of reasons. At one end of the scale, a party might ask a very high number of extremely tendentious and barely relevant requests, seeking to put the other party to enormous cost and expense in providing information of peripheral matters. At the other end of the scale, requests might be highly focused and relevant, seeking to identify matters that would provide great assistance to the case or defence being advanced by the requesting party, or to identify to the party asked to provide the information a particularly significant fact which may have been overlooked, or the significance of which has not been realised.

It is not possible to have a “one size fits all” approach to when the Authority should provide further information upon request, in any given case. However, I do not interpret the Costs Decision as applying such a test. The passages of the Costs Decision in this respect conclude that the approach and answers of the Authority to these requests, on these facts, in this case and in all the circumstances, was unreasonable. When one considers that the central issue on the reference was one of integrity, whether Mr Seiler had been reckless and lacked the necessary intention and appreciation of the Relevant Risks involved was inevitably bound into the framework within which he was working at the time. If a person is surrounded by a network of people, each or all of whom had also failed to appreciate these (or any) risks, then that may be relevant to whether that recipient of a Decision Notice also failed to appreciate those risks, or was reckless, on a case by case basis.

However, the question on this appeal is not whether this court agrees with the conclusion of the Judge on this point. The issue under consideration on this appeal on this ground is whether the view of the Judge in the Upper Tribunal was one within the range of decisions which he was entitled to make. In my judgment, it plainly was.

It must be remembered that the Authority is the regulator and able to conduct investigations, including interviewing other employees. The Judge was entitled to conclude that by taking the stance that it did, and effectively stone-walling this avenue of enquiry by Mr Seiler, the Authority had acted unreasonably. It was certainly a conclusion properly open to him in all the circumstances. I would dismiss Ground Two.

I repeat what I have said at [62] above and what Sir Stanley Burnton said at [38] in Financial Conduct Authority v Hobbs [2013] EWCA Civ 918. These proceedings are not the same as ordinary civil litigation. My conclusion on this ground does not mean that in all cases in future, unless the Authority answers all and any requests for information or clarification, it will have been acting unreasonably and will therefore automatically become liable for an adverse costs order. The regime for costs remains as it was under the Rules, with a finding of unreasonableness a threshold condition. It all must depend upon the facts.

I turn therefore to the Refused Ground, for which the Authority did not obtain permission from the single Lady Justice and which was not therefore argued before us. As I have already outlined at [2] above, in the circumstances of this case, the contents of that ground are of relevance.

This ground was in relation to what has been called the Third FX Transaction. The Third FX Transaction formed part of the Authority’s case in respect of the respondents’ conduct, and the Upper Tribunal made findings that the Third FX Transaction as advanced at the hearing before it, was significantly different from the way that it had been put in the Warning Notices issued to each of Mr Seiler and Mrs Whitestone. Additionally, the Upper Tribunal found that each of them had been prejudiced by the Authority’s change in case regarding the Third FX Transaction.

The Refused Ground was that the Upper Tribunal had erred in finding that the RDC had acted unreasonably in including findings regarding the Third FX Transaction in its Decision Notice, in the absence of any amendment to the Warning Notices, and that the Authority had conducted itself unreasonably in continuing to rely on that transaction. This means that no appeal can be brought by the Authority in respect of the finding of unreasonableness regarding the Third FX Transaction. This means that, regardless of the conclusion of this court on Grounds One and Two, the Authority would have been left, in any event, with an undisturbed and unimpeachable finding of unreasonableness in at least one major respect in its conduct of the proceedings before the Upper Tribunal.

That finding of unreasonable conduct is sufficient to clothe the Upper Tribunal with jurisdiction to make a costs order against the Authority, and exercise its discretion to do so. This is because Rule 10(3)(d) states that the Upper Tribunal may not make an order in respect of costs and expenses except “(d) if the Upper Tribunal considers that a party or its representative has acted unreasonably in bringing, defending or conducting the proceedings”. Given that “costs and expenses” must be those in respect of the proceedings, Judge Herrington could, had he so found, have made any number of different costs orders given the breadth of the proceedings which had been dealt with in the references and which had led to the findings in the substantive judgment as a whole.

It is difficult to see, given the way he approached the matter, how success by the Authority on – say – a single ground of appeal, could assist it in overturning the costs order, when the finding of unreasonableness on the Third FX Transaction would remain. His orders in respect of costs are clearly ones within the range of orders that it was properly open to him to make, based on his deep knowledge of a highly factual and complicated case, the lengthy substantive hearing, and equally highly detailed and comprehensive substantive judgment by the Upper Tribunal. As I outlined above at [6], given the length of the hearing and the highly detailed content of the substantive judgment on the reference, the Tribunal Judge would have a depth of knowledge of this matter which no appellate court could hope to match.

In my judgment, interference with his costs orders where there was an unimpeachable finding of unreasonableness in one of the three transactions before the Upper Tribunal, would be akin to unwarranted appellate interference in matters in which the Upper Tribunal was ideally placed to reach such conclusions. The degree of knowledge which this court could acquire on the matter must inevitably be a very poor second to the in-depth awareness of the proceedings, with its complex factual landscape, that the Upper Tribunal would have acquired in such a detailed case. Arriving at the conclusion that the Authority had acted unreasonably was something that the Upper Tribunal did after careful analysis of all the relevant features of the case, as can be seen from the detailed Costs Decision itself, which itself runs to 160 paragraphs. In my judgment, there are no principled grounds upon which this court should interfere with that conclusion.

Conclusion

Therefore, I would dismiss the appeal on both grounds and the costs order made below by the Upper Tribunal should remain undisturbed and in its original form.

After much thought, and with much regret, I am unable to agree with the analysis and conclusion of the majority. As my view will make no difference to the outcome, I will try to explain as succinctly as possible why I have reached it.

There are two broad points on which I disagree with the approach of the majority.

Is the decision of the Upper Tribunal in this case reviewable or not?

I accept that, in the exercise of each of its many statutory jurisdictions, the Upper Tribunal is a specialist tribunal with expert knowledge, knowledge which this court lacks. There is a countervailing point, however. The purpose of the system set up by the TCEA in the wake of the Leggatt Report (‘Tribunals for Users – One System, One Service’, August 2001) is not that each Chamber of the Upper Tribunal should be self-contained. A unifying factor is that there is an appeal on a point of law to this court from each of the jurisdictions exercised by the Upper Tribunal. The purpose of that right of appeal is to ensure that, in its various specialist areas, the Upper Tribunal does not set up its own idiosyncratic regimes, but is subject to supervision by this court so that, overall, the law is applied consistently across all tribunals.

I will only refer to three cases on this issue. The context of the first case is an appeal to the House of Lords from a decision of the Divisional Court to make a mandatory order. Section 68 of the Education Act 1944 enabled the Secretary of State to give a direction on the ground that a local authority had acted or proposed to act unreasonably in the discharge of its functions. In Secretary of State for Education and Science v Tameside MBC [1977] AC 1014 the Secretary of State for Education and Science applied to the Divisional Court for an order of mandamus requiring the respondent local authority to comply with such a direction, made on 11 June 1976. The Divisional Court granted the application but this court allowed the respondent’s appeal and the Appellate Committee unanimously upheld the decision of this court, in a decision announced on 2 August 1976. The members of the Committee gave slightly different reasons for their decisions, but they are broadly similar.

At page 1047B, Lord Wilberforce said that ‘A direction under section 68 must be justified on the ground of unreasonable action in doing what under the Act the local authority is entitled to do, and under the Act it has a freedom of choice.’ The ‘critical question’ was whether the judgment of the Secretary of State on an educational matter could be challenged. He added, at page 1047D-E: ‘If a judgment requires, before it can be made, the existence of some facts, then, although the evaluation of those facts is for the Secretary of State alone, the court must inquire whether those facts exist, and have been taken into account, whether the judgment has been made upon a proper self-direction as to those facts, whether the judgment has not been made upon other facts which ought not to have been taken into account. If these requirements are not met, then the exercise of judgment, however bona fide it may be, becomes capable of challenge: see Secretary of State for Employment v. ASLEF (No. 2) [1972] 2 QB 455, per Lord Denning MR, at p 493’. He examined the Secretary of State’s reasoning and concluded, at p 1052F-G, that the Secretary of State had ‘fundamentally misconceived and mis-directed himself as to the proper manner in which to regard the proposed action of’ the local authority’. If the Secretary of State had analysed the facts correctly, he could not have concluded that the local authority was proposing to act unreasonably.

Lord Diplock said, at pp 1064-5, ‘It is not for any court of law to substitute its own opinion for his; but it is for a court of law to determine whether it has been established that in reaching his decision…he had directed himself properly in law and had in consequence taken into consideration the matters which upon the true construction of the Act he ought to have considered and excluded from his consideration matters that were irrelevant to what he had to consider: see Associated Provincial Picture Houses Ltd v Wednesbury Corporation …Or, put more compendiously, the question for the court is, did the Secretary of State ask himself the right question and take reasonable steps to acquaint himself with the relevant information to enable him to answer it correctly?’ His conclusion, at p 1067A-B, was that the Secretary of State ‘did not direct his mind to the right question; and so, since his good faith is not in question, he cannot have directed himself properly in law.’

Lord Salmon said at p 1070F-G that to issue a lawful direction, the Secretary of State had to satisfy himself not merely that he disagreed with the authority’s proposal, nor that the authority was mistaken or wrong. The Secretary of State had to ask himself ‘Could any reasonable local authority act in the way in which this authority has acted or is proposing to act?’ He could only give a direction if there was no ‘material capable of satisfying a reasonable man’ that the answer was ‘No’. He approved a passage from In re W (An Infant) [1971] AC 682, 700, to the effect that reasonable people can disagree about a subject without being unreasonable, and that not every reasonable exercise of judgment is right, and not every mistaken exercise of judgment is unreasonable. The Secretary of State had either misdirected himself; if not, Lord Salmon could not ‘discern any valid ground upon which such a decision could be justified’ (p 1071 A-B).

An appeal from the Industrial Tribunal (now the Employment Tribunal - ‘the ET’) to the Employment Appeal Tribunal (‘the EAT’) has always been on a point of law only. An employee is unfairly dismissed if his employer acts unreasonably in treating the employer’s reason for dismissal as a sufficient reason for dismissal having regard to equity and the substantial merits of the case. The ET’s conclusion on that question has never been treated by the EAT or by this court as immune from scrutiny on appeal. An early example is the well known case of British Home Stores v Burchell [1980] ICR 303, but there are many others. An appeal may succeed if the appellant shows, for example, that the ET has misdirected itself, or taken into account irrelevant considerations, or failed to take into account relevant considerations. Equally, an appeal from the exercise of a discretion by the ET, such as the discretion to adjourn a hearing, may succeed if the ET has erred in principle, or reached a decision which no reasonable ET could have reached: see O’Cathail v Transport for London [2013] EWCA Civ 21; [2013] ICR 614, at [11], per Mummery LJ.

Obrey v Secretary of State for Work and Pensions [2013] EWCA Civ 1584; [2014] HLR 12 does not help on this question. The issue in that case was whether the Secretary of State had a justification defence to an article 14 discrimination claim. It is clear from Sullivan LJ’s judgment in that case, and in particular from paragraph 13, that the appellants simply did not acknowledge that the appeal was an appeal on a point of law and tried to persuade this court to rehear the issue of justification as if the appeal had been an ordinary appeal under CPR 52.11. This court held that the justification issue was one for the specialist assessment of the Upper Tribunal, subject only to the question whether it had erred in law in that assessment. That conclusion does not undermine what I consider to be the correct approach to this appeal.

The majority give much emphasis to the arguments that the Upper Tribunal is part of the regulatory scheme established by the legislation, that it and the FCA are engaged in a joint enterprise in the public interest, that the jurisdiction of the Upper Tribunal is supervisory, and that a reference to the Upper Tribunal in a case such as this is not ‘ordinary civil litigation’. At the risk of oversimplifying the difference between the reasoning of the majority and mine, these various formulae are treated as an automatic answer to the FCA’s criticisms of the approach of the Upper Tribunal. They are not.

In paragraph 71 of its judgment in Frensham, the Upper Tribunal described its powers on a reference.

It added, in paragraph 72, that it was well established that the burden of proof is on the FCA and the standard of proof is the ordinary civil balance of probability, that is, whether ‘the alleged events more probably occurred than not’.

The FCA, like many other administrative decision-makers, does not hear evidence from witnesses before it makes a decision. In a loose sense, the Upper Tribunal and the Administrative Court on an application for judicial review both supervise the decisions of administrative decision makers. But they do so in a different way, as is evident from a comparison between the summary in the previous two paragraphs, and the general approach of the Administrative Court. By contrast with the Upper Tribunal, the Administrative Court is restricted to the material which was before the decision-maker. Apart from the anomalous exception of age assessments, the Administrative Court does not hear evidence or make its own findings of fact. The duty of candour in judicial review is imposed because disclosure is not generally available on an application for judicial review (Tweed v Parades Commission for Northern Ireland [2006] UKHL 5; [2007] 1 AC 6). Its scope should not be exaggerated, however. It is just a duty to explain the reasons for the decision which is challenged on the application for judicial review. It is not a duty to seek out all potentially relevant evidence, or to carry out a wide factual investigation in order to help the Administrative Court with any questions it might have about the background. The extent of a decision-maker’s duty to investigate in a public law case is to do no more than to ‘take reasonable steps to acquaint himself with the relevant information’, as Lord Diplock put it in the Tameside case.

The only relevant restriction on the powers of the Upper Tribunal on a non-disciplinary reference is that, while it can hear evidence and make findings of fact, it has, in effect, to remit the question of remedy to the FCA if it considers that, on the Upper Tribunal’s findings of fact, the FCA’s decision was not open to it, unless it is inevitable that the FCA would make the same decision on the Upper Tribunal’s findings of fact. It can also make recommendations to the FCA. So, while its jurisdiction is supervisory in that limited sense, its jurisdiction is not a purely supervisory jurisdiction in the sense in which that term is used to describe the jurisdiction of the Administrative Court on an application for judicial review.

It is also true that a reference to the Upper Tribunal is not the same thing as ‘ordinary civil litigation’ under the CPR. There are many different types of litigation in the Upper Tribunal, under many different statutory regimes. Some provide for appeals, some for appeals on a point of law, and some, as in this case, for a complete rehearing. As the Upper Tribunal recognised in this case, the FCA had to prove its pleaded case in the Upper Tribunal on the evidence, on the balance of probabilities. In that broad sense, the conduct of a reference (a) does closely resemble ‘ordinary’ civil litigation, and (b) is quite different from an application for judicial review.

The majority considers that Hobbs v Financial Conduct Authority and R (Wilford) v Financial Services Authority are authorities which bind this court to take a particular approach to this appeal. They are said to show that ‘the Upper Tribunal is an expert tribunal exercising a judicial function; and that it is also an integral part of the regulatory scheme established by the Act’. That is no doubt true, but I do not understand how that description entails the consequences described at [63] of the judgment of Fraser LJ. It is necessary to unpack that consequence. It is based on two linked assumptions, which I do not share. They are that there are only two relevant types of litigation, ‘ordinary civil litigation’ and litigation in which the Upper Tribunal is ‘part of the regulatory scheme’ and that, in litigation of the second kind, the Upper Tribunal, though exercising judicial functions, is somehow immune, or virtually immune, from challenge in a case like this.

The issue in Hobbs v Financial Conduct Authority was whether proceedings had been properly discontinued by a notice on the FCA’s website, and, if not, whether the Upper Tribunal, which had held that the applicant was not guilty of market abuse, had erred in law in not finding for the FCA on the ground that the applicant was, nevertheless, not a fit and proper person. The point which Stanley Burnton LJ was making in paragraph 38 of his judgment, when he said that a reference was not ‘ordinary civil litigation’ was simply that, on a reference, the Upper Tribunal had to consider all the facts and evidence that were put before it at the hearing. His point was that there had been two limbs to the FCA’s case against the applicant and that the Upper Tribunal should not have confined its consideration to only one of those limbs. His reasoning goes no further than that.

The issue in Wilford was whether the Administrative Court had erred, on an application for judicial review, in quashing a decision of the FCA, when the claimant had a suitable alternative remedy, that is, a statutory right to refer the matter to the Upper Tribunal. Paragraph 21 of the judgment of Moore-Bick LJ makes the self-evident point that the Upper Tribunal is ‘part of the regulatory scheme’. It casts no light on how the Upper Tribunal should approach its judicial functions on a non-disciplinary reference, however, still less does it help with the question whether this court can intervene in a case like this. Paragraph 37 of Wilford has no bearing on this issue either, despite the hint at [66] of Fraser LJ’s judgment that it might have. Paragraph 37 of Wilford simply emphasises that judicial review is not usually available when there is a suitable alternative remedy. Moore-Bick LJ was not saying that the Upper Tribunal is immune from scrutiny by the court, but rather that, as an applicant can take his case to the Upper Tribunal, the Administrative Court should not get involved in scrutinising the activities of the FCA.

Those paragraphs of those judgments do not exclude the intervention of this court in an appropriate case. Neither Stanley Burnton LJ nor Moore-Bick LJ was considering that issue, and neither passage in either judgment binds this court on this issue. The public interest which is undoubtedly engaged by the FCA’s activities and by the right of an applicant to refer a decision to the Upper Tribunal is a relevant background feature, but it does not decide this issue.

For those reasons, I would hold that this court can scrutinise the reasoning and conclusion of the Upper Tribunal in the exercise of its power to award costs, and that it should allow the appeal if it holds that the Upper Tribunal has materially erred in law (in the sense which I have described). That answer is not affected by the fact that the ‘threshold condition’ for the exercise of the power to award costs is a finding that the FCA has acted unreasonably (see the Tameside case). I also accept that there is a difference between the test which is applied on an ordinary appeal: that is, whether the decision of the judge is ‘plainly wrong in the sense that it was not reasonably open to him’ and the Wednesbury test, which applies to an appeal of a point of law. That latter is a somewhat stricter test for intervention (see per Lord Fraser of Tullybelton in G v G [1985] 1 WLR, at pp 653E-F, whom the other members of the Appellate Committee agreed).

Did the Upper Tribunal materially err in law?

The starting point is three relevant findings of the Upper Tribunal.

At [48] of his judgment Fraser LJ cites paragraphs 86-89 of Frensham. The facts of that case explain why the Upper Tribunal said what it said in those paragraphs. The applicant in that case was convicted by a jury of a sexual offence. It had no obvious connection with his regulated activity. He had also breached the conditions of his bail. He was sentenced to 22 months’ imprisonment, suspended. The FCA decided to withdraw his approval pursuant to section 63 of the FSMA and to make a prohibition order on the grounds that he lacked integrity. The FCA also considered that the applicant had not been transparent with it, as he had not told the FCA of his earlier arrest, his later arrest and remand in custody, that he had not been able to discharge his controlled functions for five weeks or ensure the compliance of his firm with its regulatory obligations, and that the Chartered Insurance Institute had both refused to renew his statement of professional standing and expelled him.

The applicant’s case was that the FCA had taken irrelevant considerations into account. The findings that he lacked candour were wrong. He was ‘deeply remorseful’ and the sanction was ‘disproportionate’. The Upper Tribunal noted that this was the first case in which it had had to consider a challenge to a prohibition order which had been made after a conviction for an offence which did not involve dishonesty and when the relevant conduct was not concerned with the applicant’s regulated activity. I will refer to that as ‘an extraneous conviction’.

The Upper Tribunal was, understandably, perhaps, interested in the history of the FCA’s policy about extraneous convictions. The FCA’s witnesses in the Upper Tribunal were Ms Couzens and Mr Blunt. They knew next to nothing about that issue. In paragraph 80, the Upper Tribunal said that it was ‘not helpful’ that the FCA had failed to call witnesses who knew about the policy. Further, the FCA had not been appropriately candid about the reasons for its delay in making a decision in the case (paragraphs 81-84). The FCA appeared not to want to tell the Upper Tribunal that its policy about extraneous convictions had only been adopted ‘after a long period of reflection’ (paragraph 86).

Paragraph 88 is about disciplinary references, not non-disciplinary references, but it recognises that proceedings in the Upper Tribunal are largely adversarial. Paragraph 89 is about non-disciplinary references. It makes the point that the tribunal needs to ‘understand the procedural and other steps taken in connection with the making of’ the FCA’s decision. For the reasons I have already given the similarities between judicial review and a non-disciplinary reference are very limited, contrary to the view of the Upper Tribunal. I nevertheless accept that because the FCA is public body governed by public law, it must be candid with the Upper Tribunal. But the scope of that duty of candour can be no greater than the duty of candour in judicial review, that is, a duty to explain its decision-making to the Upper Tribunal. That duty could, therefore, in an appropriate case, require the FCA to explain the development of a relevant policy to the Upper Tribunal, in order to enable the Upper Tribunal to understand the relationship between the FCA’s functions, the public interest and the relevant policy. The policy about extraneous convictions in Frensham is an example. The last sentence of paragraph 89 must be understood in the context of the facts of Frensham; in other words, what the Upper Tribunal expected was a full and accurate explanation of matters in the FCA’s own knowledge relating to the process by which the decision was made: of the development of, and reasons for, any relevant policy, and, where relevant, of any otherwise unexplained delays in the process.

I accept that the FCA and the Upper Tribunal are engaged in a ‘common enterprise with the Upper Tribunal in ensuring that the objects of the legislation are achieved’ and public confidence is maintained (at [51] of the judgment of Fraser LJ). Section 133A(5) of the FSMA may well support that view, and the view that a non-disciplinary reference is not ‘ordinary civil litigation’. But for the reasons I have already given, those considerations are not an answer to the question whether the Upper Tribunal’s decision on costs can be challenged; nor, contrary to the apparent suggestion at [51] above, are they an answer to the FCA’s criticisms of the Upper Tribunal’s approach to costs in this case.

For the reasons I have already given the assertion that whether or not conduct is unreasonable is a question of ‘evaluative fact’ it is not necessarily an answer to those criticisms. That assertion ignores the possibility that, in making that evaluation, the Upper Tribunal, as the Secretary of State in Tameside did, has misdirected itself, or reached a decision which was not properly open to it (that is, a decision which is Wednesbury unreasonable).

I will now consider the Upper Tribunal’s views about the FCA’s conduct in relation to witnesses, and then its response to Mr Seiler’s requests for further information.

The first question is what the Upper Tribunal found. In paragraph 116 of the costs decision, it summarised Mr Seiler’s submissions. He submitted that the FCA unreasonably failed to call material witnesses: Mr Campeanu and Mr Narrandes, and that it had not even explained whether it had considered asking Ms Thomson Bielmann and Mr Courrier for help, even though their evidence would have been ‘highly relevant’ (my emphasis). He also submitted that the FCA had acted unreasonably in the way it responded to his request for further information, and had acted unreasonably in not telling him more about the details of its investigation.

The Upper Tribunal accepted those submissions in relation to Mr Seiler’s case. In paragraph 118, it said that ‘As far as the absence of witnesses is concerned, there is no question that the Tribunal would have been assisted by evidence from’ those four people. It criticised the FCA for not considering whether to ask it for directions about Mr Campeanu (whom the FCA did not regard as a witness of truth), for failing to consider whether or not to call Mr Narrandes, when his absence as a witness was ‘unsatisfactory’, for not telling the Upper Tribunal at the substantive hearing about its contacts with the other two witnesses, which was ‘totally unacceptable’, for its ‘utterly feeble attempts’ to get their help, for sending letters to them which encouraged a negative answer, and for pretending to be serious about getting their help. The FCA ‘took no steps to demonstrate to the Tribunal that it has tried to obtain all the relevant evidence’ (paragraph 124 of the costs decision) (my emphasis again). It did not consider whether to ask the Upper Tribunal to issue letters of request. Its conclusion was that the FCA had acted unreasonably ‘in the conduct of the proceedings in relation to its approach to potential witnesses referred to above’ (paragraph 128). In paragraph 129, it held that the FCA’s response to Mr Seiler’s request for information was also unreasonable. It accepted his submission that it would have been ‘assisted’ if the FCA ‘had given a more positive response’.

In paragraph 145, the Upper Tribunal recorded that Ms Clarke had made similar submissions for Ms Whitestone. It decided to make an order for costs in her case ‘in relation to the matters referred to at paragraph 116’.

There was only one issue about each applicant. The FCA’s case was that each applicant lacked integrity on the sole ground that he or she recklessly disregarded risks of which they were actually aware that Julius Baer might be facilitating or taking part in financial crime (paragraph 6 of the costs decision). The applicants’ case was that they did not know about the risks (paragraph 7). The UT rejected the FCA’s case (paragraph 10). It decided, in short, that the applicants were incompetent (paragraphs 12-14). It is important to bear in mind that, in the costs decision, the Upper Tribunal rejected the primary case of both applicants. It held that the FCA had acted reasonably in making the decisions in each case, and that, with specific exceptions, the FCA acted reasonably in bringing, defending or conducting the proceedings. That was because, as regards the decision, on the available documentary evidence, the FCA could reasonably have concluded that the applicants did lack integrity in that way, and, as regards the proceedings, because it was only after prolonged examination of their oral evidence (something which the FCA, as administrative regulator, does not and did not do) that the Upper Tribunal was able to make the findings of fact in each case that the explanation for the applicant’s conduct was that he or she was incompetent, and not that he or she lacked integrity. Paragraph 114 of the costs decision explains this point in the case of Mr Seiler (and see paragraphs 131-143 for Ms Whitestone).

The Upper Tribunal referred to the evidence of Mr Neary, for the FCA, accepting that ‘it is relevant to look at how other people reacted at the time’ (costs decision, paragraph 129). In paragraph 153 of the substantive decision, the Upper Tribunal recorded that Mr Neary had accepted in his evidence that other people’s reaction is relevant to an assessment of the applicants’ integrity. Mr Neary’s evidence was not factual evidence. He was expressing an opinion. With all respect to Mr Neary, that opinion is plainly wrong. Whether a person lacks integrity for the purposes of a regulatory framework is an objective question, to be decided (as indeed the Upper Tribunal did decide it: see above) on the basis of detailed findings of fact about the person’s subjective knowledge about what was, or was not, going on in the course of his or her regulated activity.

This approach to relevance coloured the Upper Tribunal’s findings about the evidence of the four potential witnesses, and its findings about the FCA’s response to Mr Seiler’s requests for information. In this respect, the Upper Tribunal misdirected itself. However interesting it might have been for the Upper Tribunal to hear evidence from the four potential witnesses, and however interesting it would have been for the FCA to respond more fully to Mr Seiler’s requests than it did, neither that evidence, nor those requests, were legally relevant to the issues which the Upper Tribunal had to, and did, decide. Moreover, these potential witnesses were in a different category from the witnesses whom the FCA failed to call in Frensham. They were not employees who knew about the development of the FCA’s relevant policy or about the linked question of delay. They were potential witnesses of fact, who were not employed by the FCA, and whose evidence was not relevant to the real issues in the case.

There are two further errors of principle in the Upper Tribunal’s approach to the evidence from the four potential witnesses. The first is linked with the Upper Tribunal’s mistaken view that their evidence was legally relevant. If a party to a reference fails to produce legally relevant evidence in support of its pleaded case, in circumstances where it could have called the witness, it may well fail to prove its case, ‘lose’ the reference, and may, in appropriate circumstances, be found to have acted unreasonably for the purposes of costs. But it is wrong in principle to hold that a party has acted unreasonably in failing to call evidence which it considers is, and which on analysis is, legally irrelevant to its pleaded case. The second error is also linked with the Upper Tribunal’s mistaken view of the scope of the FCA’s duty of candour. That duty does not extend to co-operating with the Upper Tribunal either by calling a witness whom the FCA considers is not a witness of truth, and/or by not calling witnesses on whose evidence the FCA does not need to rely in order to prove its pleaded case because it is evidence which is legally irrelevant to that pleaded case. That duty does not impose an obligation to call people who had some connection with, or some knowledge of, the events at issue. Compliance with a supposed duty to call peripheral witnesses is liable simply to increase costs for both sides, and unnecessarily to increase the length of the hearing. It is inimical to the overriding objective and is the epitome of unreasonable conduct. If the applicants thought that these people would help their case, it was always open to them to call them as witnesses, and to ask the Upper Tribunal for a witness summons if they did not co-operate. In the event, the applicants succeeded without their evidence, perhaps because their evidence was not relevant to the real issues.

For these purposes I am not persuaded that there is a legally relevant distinction between criticising the FCA for not calling a witness and criticism of the FCA for not taking steps falling short of calling a witness (such as asking the Upper Tribunal for directions about his evidence, considering whether or not to call him, or making timely and sincere efforts to seek his or her co-operation well in advance of the hearing date). If the evidence of the potential witness was not legally relevant, it was not open to the Upper Tribunal to find that the FCA acted unreasonably in not calling that witness. A fortiori, the FCA could not have acted unreasonably in not taking steps falling short of calling that witness, or taking such steps, but insincerely. In other words, the case for criticism of the FCA is weaker, not stronger, in relation to steps falling short of calling a witness.

I also consider that the Upper Tribunal misdirected itself about Mr Seiler’s requests for information. The requests were about legally irrelevant matters, and, in my view (for what it is worth) were properly answered by the FCA. Whether or not that is so, there is a further distinct point, which is a complete answer to this criticism. If Mr Seiler, as it seems that he was, was dissatisfied with the FCA’s responses to his requests, the Tribunal Procedure (Upper Tribunal) Rules 2008 (‘the Rules’) provided the remedy. He should have applied to the Upper Tribunal for a direction requiring the FCA to comply with his requests, and if the FCA did not comply, he could, ultimately, have asked the Upper Tribunal to strike out the FCA’s case (see rules 6 and 8 of the Rules). The Upper Tribunal was wrong in principle, or itself acted in a way which was Wednesbury unreasonable, in deciding that the FCA should be liable in costs, after the event, when Mr Seiler had not himself, in the course of the proceedings, invoked those two remedies given by the Rules in relation to the FCA’s responses to his requests for further information. Those rules provided the appropriate discipline and sanction, not a later application for costs.

[92]-[97] of the judgment of Fraser LJ give reasons why, even if the Upper Tribunal erred in law in relation to the two grounds of appeal for which permission to appeal was given, its award of costs should nevertheless be upheld, in short, because the FCA was refused permission to appeal in relation to the ground of appeal based on the third FX transaction. I do not agree with this reasoning, on two grounds. First, it cannot be said (if it is assumed that the Upper Tribunal erred in law as argued in the two other grounds of appeal) that those errors of law were not material. In other words, it is impossible to say that, had the Upper Tribunal not made those errors, it would inevitably have made the same costs award based on the third FX transaction alone. Second, this point was not relied on by the Upper Tribunal. It is a point for a Respondent’s Notice. There was no Respondent’s Notice on this appeal. I would allow the appeal.

I have read the judgment of Fraser LJ in draft; and I agree with it. I make these short observations in the light of Elisabeth Laing LJ’s disagreement.

The threshold condition for the exercise of the Upper Tribunal’s discretion to award costs is:

There are two parts to this condition. The first is that the Upper Tribunal forms an opinion (“the Upper Tribunal considers”). The second is that its opinion is that person has conducted proceedings unreasonably. The Upper Tribunal’s discretion is not engaged at this stage.

Whether someone has or has not conducted proceedings unreasonably is an evaluative factual judgment. How does a challenge to such a factual evaluation fit into a regime where the permitted grounds of appeal are limited to a question of law? The answer, in my view, is given by Edwards v Bairstow [1956] AC 14. In his classic exposition of what amounts to an error of law, Lord Radcliffe said:

I would not, therefore, accept that the majority decision means that what appears to be a finding of fact by the Upper Tribunal is unreviewable.

Whether proceedings have been conducted unreasonably is, to some extent at least, a comparative exercise. It is the Upper Tribunal, and not this court, which has extensive experience of how in practice references such as this have been (and ought to be) conducted. The evaluative judgment of the Upper Tribunal on a question like this commands the utmost respect from this court.

In addition, in a highly specialised area like this one, one of the primary functions of the Upper Tribunal is to lay down standards of practice without undue interference by generalist appellate courts: compare R (Jones) v First Tier Tribunal (Social Entitlement) Chamber [2013] UKSC 19, [2013] 2 AC 48 (Lord Carnwath); AH (Sudan) v Home Secretary [2007] UKHL 49; [2008] 1 AC 678 (Lady Hale).

I am not willing to substitute my own opinion for that of the highly experienced Upper Tribunal in a case like this one.