BAILII [Home] [Databases] [World Law] [Multidatabase Search] [Help] [Feedback] [DONATE]

Scottish Court of Session Decisions

THE FUTURE OF BAILII DEPENDS ON USERS LIKE YOU


If you want to be able to use BAILII in the future, please consider making a donation to celebrate BAILII's 25 years of providing free access to law.


Your donation, no matter the size, will help BAILII maintain the legal databases that you and many other users rely on. If every visitor this month gives just £5, it will have a significant impact on BAILII's ability to continue providing this vital service.


Thank you for your support!



You are here: BAILII >> Databases >> Scottish Court of Session Decisions >> The Advocate General for Scotland for an order under section 6(4) of the Insolvency Act 1986 (Court of Session) [2026] CSOH 29 (25 March 2026)
URL: https://www.bailii.org/scot/cases/ScotCS/2026/2026csoh29.html
Cite as: [2026] CSOH 29

[New search] [Context] [View without highlighting] [Printable PDF version] [Help]


OUTER HOUSE, COURT OF SESSION
[2026] CSOH 29
P241/26
OPINION OF LORD SANDISON
in Petition of
THE ADVOCATE GENERAL FOR SCOTLAND, for and on behalf of the
COMMISSIONERS FOR HIS MAJESTY'S REVENUE AND CUSTOMS
Petitioner
for
an order under section 6(4) of the Insolvency Act 1986 revoking the decisions of the creditors
and shareholder of Petrofac Facilities Management Limited dated 30 January 2026
approving a Company Voluntary Arrangement
Petitioner: Ower, KC, Reid; Office of the Advocate General
Respondent: Borland, KC, Steel; Burness Paull LLP
25 March 2026
[1]
By this petition, the Advocate General on behalf of HMRC asks the court to order
in terms of section 6(4)(a) of the Insolvency Act 1986 that the decisions of the creditors and
member of Petrofac Facilities Management Limited ("PFML") on 30 January 2026 approving
a Company Voluntary Arrangement ("CVA") should be revoked. PMFL is the respondent
to, and opposes, the petition, and given its extremely precarious financial state and the
apparent need for the sale transaction underlying the CVA to be completed quickly, sought
its expedited disposal. Parties lodged affidavits (on the part of the company from the chief
financial officer of the Petrofac Group and one of the joint supervisors of the CVA, and
2
on the part of HMRC from a relevant compliance manager) to deal with and support the
factual elements of their respective positions. A hearing took place on 12 March 2026,
within 2 weeks of the commencement of proceedings, at the conclusion of which I refused
the prayer of the petition. This opinion provides the reasons for that decision.
CVAs
[2]
A CVA is a procedure for restructuring certain debts of a company in financial
distress, governed by Part I of the 1986 Act, read together with the Insolvency (Scotland)
(Company Voluntary Arrangements and Administration) Rules 2018 (SSI 2018/1082).
Section 1(1) of the 1986 Act defines a CVA as "a composition in satisfaction of [the
company's] debts or a scheme of arrangement of its affairs". A resolution to approve a
CVA is put to the company's creditors and shareholders. Approval is given if 75% or more
by value of those responding vote in favour of it, so long as no more than half of the total
by value of creditors not connected with the company vote against it. Once approved, a
CVA is binding on all the creditors of the company (subject to certain minor exceptions),
including those who voted against it: section 5 of the 1986 Act. Creditors vote on a CVA
without being divided into different voting classes, even if they are treated differently under
its terms. The CVA procedure is a flexible process designed to attempt to assist a company
in financial distress and to prevent it from entering into formal insolvency proceedings. It
is part of the corporate "rescue culture" described in the Cork Report (Report of the Review
Committee on Insolvency Law and Practice (1982) Cmnd 8558). A CVA cannot affect the
rights of a secured creditor in respect of its security: section 4(3) of the 1986 Act. However,
where the security is worth less than the total amount of secured debt, the shortfall is treated
as an unsecured claim on which the relevant creditor can vote. In addition, any debts
3
ranking as preferential debts (including certain tax liabilities) can only be compromised to
a very limited extent by a CVA: section 4(4) of the 1986 Act.
PFML and the Petrofac Group
[3]
PFML is part of the Petrofac Group, a construction and engineering contractor for
major energy projects around the world. The ultimate parent company of the group is
Petrofac Limited. The group has three main business divisions: (i) Engineering and
Construction; (ii) Asset Solutions; and (iii) Integrated Energy Services. PFML is the main
contracting entity and employer within the Asset Solutions division. That division manages
and maintains onshore and offshore energy operations, as well as decommissioning services
for obsolete infrastructure, and employs approximately 3,000 people (over half of whom are
directly employed by PFML), predominantly in Scotland. The division also has hundreds
of suppliers within its supply chain, mostly in Scotland, and in particular in Aberdeen.
[4]
The respondent avers, and the petitioner does not seek to deny, that the group
has been in financial distress for several years. It carried out a debt restructuring exercise
in 2021, but its financial position continued to worsen. Its distress has a range of causes,
including the lasting impact of the COVID-19 pandemic, severe supply chain disruption
due to the Russian invasion of Ukraine in February 2022 and associated costs in connection
with the group's exit from Russia; and factors specific to the Engineering and Construction
division, including costs overruns of more than US$1 billion on a project in Thailand.
In 2023, the group commenced further restructuring negotiations with its creditors. In
December 2024, Petrofac Limited and one of its subsidiaries proposed a restructuring plan
under Part 26A of the Companies Act 2006 in the High Court of England and Wales. The
plan was designed to restructure the financial liabilities of many companies within the
4
group, and thereby to restore it to financial health. On 20 May 2025, the plan was sanctioned
at first instance by the High Court of England and Wales: [2025] EWHC 1250 (Ch).
However, on 1 July 2025, that judgment was reversed by the Court of Appeal: [2025] EWCA Civ 821,
and the order sanctioning the plan was set aside. The group was thus left on the
brink of collapse. A further attempt to effect a group-wide restructuring was undertaken
(based on a pre-packaged administration sale of the vast majority of Petrofac Limited's
subsidiaries). That transaction was due to complete in November 2025. However, in
October 2025 a key customer terminated its contracts with the group, rendering that
turnaround plan unviable. Petrofac Limited also became cashflow insolvent. In those
circumstances, its directors had no choice but to file for administration under Schedule B1
to the 1986 Act. On 27 October 2025, the High Court of England and Wales made an
administration order in respect of it. Certain other companies within the group have
subsequently collapsed, or are expected to shortly enter into administration and other
formal insolvency proceedings around the world. Nevertheless, most of Petrofac Limited's
underlying operating subsidiaries, including PFML, have continued to carry on business.
Immediately after their appointment, Petrofac Limited's administrators launched an
expedited sale process in respect of the group's Asset Solutions division, in which PFML
is the main trading entity. The administrators engaged with over forty trade and financial
parties as part of the sales process. Following a competitive process which generated nine
phase one bids and three phase two bids, the administrators selected a US company, CB&I
(originally known as Chicago Bridge & Iron Co) as the preferred bidder. On 24 December
2025, Petrofac Limited and certain other group entities entered into a sale and purchase
agreement with various CB&I entities, pursuant to which the buying entities would inter alia
acquire the share capital of PFML. The completion of that transaction is subject to several
5
fundamental conditions precedent, including a full and final settlement of the dispute with
HMRC shortly to be described and a release of the great majority of PFML's other liabilities,
including the claims of PFML's secured creditors (amounting to more than £1 billion). It
is claimed that the buyers would be unwilling to enter into the sale transaction without a
compromise of those liabilities.
PFML and HMRC
[5]
There is a long-running and aged tax dispute between PFML and HMRC. HMRC
claims that PFML adopted a wholly artificial structure for the purpose of avoiding liability
to secondary Class 1 NICs in respect of workers employed by two Jersey companies in the
group for the period from October 1999 to April 2014. It claims to be owed £151,062,514
(consisting of principal in the sum of £91,793,949 and interest in the sum of £58,807,142), on
the basis that PFML was a "host employer" under paragraph 9 of Schedule 3 to the Social
Security (Categorisation of Earners) Regulations 1978. Those Regulations were amended
in 2014, and the affected employees returned to PFML in 2016. PFML disputes that the
legislation justifies the claim, which is the only one against it by HMRC that would be
compromised by the CVA. Ostensibly similar arrangements entered into by other
companies in the oil industry have been found to be ineffective to avoid NIC liabilities,
but the dispute with PFML has not yet been judicially determined. In January 2025, as one
element of the Part 26A restructuring plan already described, HMRC signed a settlement
agreement with PFML which provided for the payment of £22.5m over a period of 5 years
and 6 months. That agreement was arrived at based on support to be provided to the group
by an ad hoc assembly of secured creditors and noteholders. It was contingent upon the
6
sanction of the Part 26A plan; as already narrated, that plan did not proceed and the agreed
settlement fell with it.
[6]
The Assets Sale Agreement which has now been reached with CB&I requires the
compromise of HMRC's claim against PFML, along with various other liabilities. Against
that background, a proposal for a CVA was made on 14 January 2026. PFML's creditors
and its sole member passed resolutions in favour of the proposal on 30 January. It is that
decision which is the subject of the present petition. HMRC's relevant claim is neither
secured nor preferential, and so is in principle one with which a CVA may properly deal.
The terms of the CVA
[7]
Under the CVA, the creditors of PFML are divided into nine categories (labelled A
to I), although these categories are not voting classes, since all creditors vote together
when deciding whether to approve a CVA. The first five categories (A to E) comprises
the creditors whose claims will be compromised and released by the CVA, in exchange
for a pro rata share of a fund derived from the proceeds of the sale transaction, calculated
in accordance with a formula set out in the CVA. HMRC's claim is treated in Category A.
It is dealt with in the same way as Category B, which consists of substantial unsecured
liabilities owed to HSBC; Category C, which consists of historical unpaid fees owing to
various professional advisers, including Moelis; Category D, which consists of certain
intra-group liabilities; and Category E, which consists of certain liabilities arising out of the
Engineering and Construction division, that does not form part of the sale transaction. The
next three categories (F to H) consist of liabilities that will not be compromised by the CVA
but will be compromised as part of the wider sale transaction of which the CVA forms part.
Categories F and G consist of PFML's secured liabilities, amounting to approximately
7
£1.052 billion. The claims of the secured creditors against PFML will be released in their
entirety after receipt of a share of the proceeds of the sale transaction amounting to
approximately 4% of the secured creditors' total claims. The release will not be given effect
by the CVA itself, since a CVA cannot affect the rights of the secured creditors in respect
of their security. Instead, it will be given effect pursuant to an inter-creditor agreement
binding on the secured creditors. The release will take place inter-conditionally with the
CVA. Category H consists of certain intra-group liabilities which will be given up in return
for nominal consideration, with certain low-value exceptions which cannot be transferred
for such consideration for reasons relating to tax or local law. The final category (I) consists
of ordinary unsecured claims that will not be compromised by the CVA, principally
(although not exclusively) because the relevant creditors are critical suppliers of goods
and services to PFML. Under the CVA regime, insolvency practitioners are required to be
appointed as "nominees" (overseeing the CVA proposal and voting) and as "supervisors"
(overseeing the implementation of the CVA). In the present case, the CVA nominees and
supervisors were and are experienced insolvency practitioners from Teneo.
Submissions for the petitioner
[8]
On behalf of the petitioner, senior counsel noted that section 6(1)(a) of the 1986 Act
provided that an application might be made to the court on the basis that a voluntary
arrangement which had effect under section 4A unfairly prejudiced the interests of a
creditor, member or contributory of the company. If the court was satisfied that the unfair
prejudice test was met, it could revoke the decision approving the CVA. The concept of
unfair prejudice was not given a more detailed definition in the statute. There was no single
and universal test. It had been said that the court might assess matters from different
8
angles, including using what were called "vertical" and "horizontal" comparisons:
Prudential Assurance Co Ltd v PRG Powerhouse Ltd [2007] EWHC 1002 (Ch),
[2007] Bus LR 1771, [2007] BCC 500 per Etherton J at [747].
[9]
The vertical comparison was a comparison between the position contained in the
CVA and the position in a realistically available alternative process, which might be
winding up, but might be some other process, depending on the facts of the case: Discovery
(Northampton) Ltd v Debenhams Retail Ltd [2019] EWHC 2441 (Ch), [2020] BCC 9 per Norris J
at [12]. That was the irreducible minimum below which the return in the CVA could not go.
In other words, if the challenging creditor would be worse off in the position contemplated
by the CVA than in the vertical comparator scenario, the challenge would be likely to
succeed: Mourant & Co Trustees Ltd v Sixty UK Ltd [2010] EWHC 1890 (Ch), [2010] BCC 882
per Henderson J at [67(c)]. The vertical comparison thus set a lower limit below which the
CVA could not go; passing that limit did not, in itself, result in the failure of any challenge:
determinative: Prudential Assurance at [83].
[10]
The horizontal comparison was a comparison between the challenging creditor's
treatment under the CVA and the treatment thereunder of other creditors or classes of
creditors. It was accepted that a CVA might involve differential treatment of creditors or
classes of creditors. The question was whether that treatment was unfair to HMRC in the
specific circumstances of the case. It was too narrow to ask simply whether the creditor
making the challenge would be better off than in an insolvency situation: Re a debtor (No 101
of 1999) (No 1) [2001] 1 BCLC 54 per Ferris J at 63 - 64, followed in Prudential Assurance
at [84] - [85]. It was relevant to ask whether the proposed arrangement was one in which
the compromise was achieved by compromising one creditor's debts by way of the votes of
other, uncompromised, creditors who would stand to benefit. Although a CVA passed by
9
the votes of uncompromised creditors was not inherently prejudicial, that was not
conclusive of the question of whether or not it should be upheld. That depended on all
the circumstances of the case. In Lazari Properties 2 Ltd v New Look Retailers Ltd
[2021] EWHC 1209 (Ch), [2021] Bus LR 915, [2022] 1 BCLC 557, it was submitted that the CVA
which had been agreed there would not be prejudicial provided that any (horizontal)
differential treatment was objectively justified, the vertical comparator was satisfied, and
a reasonable and honest person in the position of the applicant would have approved the
CVA. That submission was rejected by Zacaroli J. An important consideration was whether
there was a fair allocation of the assets available within the CVA between the compromised
creditors and other groups of creditors. That would include considering the source of the
assets from which the treatment of the different groups derived, and whether they would
or could have been made available to all creditors in the relevant alternative. It might be
relevant to ask whether the combined voting power of uncompromised creditors had been
used to remove a competing debt at no cost to them: Re a debtor (No 101 of 1999) at 62 - 63.
Similarly, whilst the test of the reasonable and honest man in the same position as the
applicants was a relevant factor (and included a comparison with how other creditors
similarly affected by the CVA had voted), that test should not be applied in too a rigid
fashion: SISU Capital Fund Ltd v Tucker [2005] EWHC 2170 (Ch), [2006] BCC 463
per Warren J at [76] - [77]; Lazari at [198]; Re T&N Ltd [2004] EWHC 2361 (Ch),
[2005] 2 BCLC 488 at [81]. Where the votes of creditors in favour of the proposal were put forward
as an example of approval by the reasonable and honest man, it was relevant to ask what the
wider interest of these creditors was and how it might differ from the objecting creditor. In
Re a debtor (No 101 of 1999), Ferris J held that it was a relevant consideration that HMRC
would retain its rights, including the ability to press for a more satisfactory arrangement.
10
Some other first instance decisions in England (in particular SISU Capital Fund) suggested
that the required comparison was between the proposed compromise and no compromise
at all, but there was no proper basis to read such a restriction into section 6(1)(a) of the
1986 Act. That assessment ought to be a broad one and should take account of the position
which the compromised creditor was losing.
[11]
The petitioner submitted in summary that the CVA was unfairly prejudicial to
HMRC's interests. The "irreducible minimum" of the vertical comparison appeared likely
be met (albeit this was contingent on completion of the underlying Assets Sale Agreement),
but the difference with an insolvency situation was relatively minimal. The CVA, however,
did not pass the horizontal comparison. It was an agreement which provided an advantage
to other groups of creditors, either by the non-compromise of their debts or by advantages
which would accrue to them from the broader Assets Sale Agreement. The votes of these
creditors had drowned out HMRC, at whose expense the transaction had been effected. It
followed that the CVA was unfairly prejudicial.
[12]
Looking first at the vertical comparison, that was between the CVA proposal on
the one hand and the unsecured creditors' share of the prescribed part in an administration
or liquidation on the other. The fund for compromised creditors for which the CVA made
provision would provide HMRC with a minimum return of 0.45% and was estimated to
provide it with a return of 2.23%. By contrast, HMRC's prescribed part in insolvency had
an estimated dividend return of 0.12%. On the face of those figures, the CVA would satisfy
the vertical comparison. As already noted, however, that was not determinative of whether
the proposal was unfairly prejudicial, but rather was merely an indication that it did not
fail by reason of not meeting an irreducible minimum. The petitioner observed that the
difference between the two potential outcomes was a relatively small one, particularly when
11
the scale of HMRC's claim was considered, and that any payment to be made out of the
fund for creditors compromised by the CVA was subject to two contingencies. That
payment would only be received at all if the Assets Sale Agreement completed. That was
subject to the satisfaction of 17 conditions precedent. If the Assets Sale Agreement did not
complete, the payment due to the fund would require to be calculated in accordance with
Clause 15 of the CVA. The final sum was therefore unknown (subject to the provision of
a 0.45% minimum). Nothing more than that was assured.
[13]
Turning to the horizontal comparison, the total value of the creditors opposing
the CVA was £152,156,001.27 (principally composed of HMRC's claim of £151,798,673).
Breaking down the votes by value of the claims of the creditors voting, the total value of
all voting creditors' claims was £1,083,592,484.54. Creditors with claims of £925,594,199.75
voted in favour of the CVA. Of the creditors voting for the CVA, £742,768,058 were in
respect of claims not compromised by it. That is, 80.25% of all creditors by value voting in
favour of the CVA were uncompromised creditors. In addition, connected creditors in
Category D voted for the claim. The value of their claims was £127,419,419.
In total, £870,187,477 of creditors, by value, voting in favour of the CVA (94.01% of the total
voting in favour) were either uncompromised or connected.
[14]
The respondent maintained that, as a matter of principle, it was not an automatic
source of prejudice that the CVA was passed by the votes of uncompromised creditors.
So far as it went, that was a correct general proposition. It remained necessary, however,
to consider in each case whether the drowning out of a compromised creditor was unfair.
Further, the claims of the Category F and G senior secured creditors were claims in relation
to guarantees of other group debt given by PFML. It appeared that those creditors had
compromised their claims as part of the broader sale transaction. Their total recovery was
12
estimated at 10 - 20%. That recovery was generated because the entire proceeds of the
sale transaction, less the fund to be made available to creditors whose claims were to be
compromised by the CVA and the costs of the CVA, would be remitted to those senior
secured creditors.
[15]
The respondent pointed out that the CVA was approved by two other unsecured
creditors whose claims would be compromised, HSBC and Moelis. However, approval by
those creditors (whose claims were in different categories to HMRC's, B and C respectively)
could not in itself be conclusive of the analysis of HMRC's position. In any event, the claims
were not comparable. They were for markedly lesser amounts (£47,978,452.49 - HSBC;
and £7,354,511.43 - Moelis). PFML's liability to HSBC was as a guarantor. The sum sought
was also sought from Petrofac International Limited. Moelis were engaged in 2025 by the
group to solicit offers from third parties (apparently, as far as PFML was concerned, with at
least a partial success fee component). The respondent estimated the ultimate shortfall that
would be suffered by HMRC and Moelis would be 80 - 90%. HMRC's shortfall would be
of the order of 97.7%. The respective claims could not be said to be comparable, and in
any event the difference in recovery as between those creditors was a striking one.
[16]
It was further maintained by the respondent that HMRC's claim had been disputed
for 14 years. However, the valuation of HMRC's claim, along with the valuation of other
claims, was a matter for the chair of the creditors' meeting, whose judgment on the matter
was not criticised by the respondent. In any event, HMRC's decision on PFML's historic
NIC liability was issued in 2020 and was thereafter appealed by PFML to the First-tier
Tribunal (Tax Chamber). Those proceedings remained ongoing, a substantive hearing listed
for February 2025 having been discharged owing to the proposed Part 26A arrangement in
relation to Petrofac Limited. PFML's appeal was the last in a series of five appeals, each
13
concerned with arrangements in the offshore industry of substantially the same character.
In each case, the taxpayer had been unsuccessful, albeit appeals to the Upper Tribunal were
pending in some cases.
[17]
Comparison between the position of HMRC and the Category D connected creditors
was also misplaced. Whether or not weight was given to those votes, it was clear that
HMRC's vote was overwhelmingly drowned out. In any event, however, the connected
nature of the Category D creditors (other companies within the Petrofac Group) was plainly
a relevant consideration. Indeed, it was a statutory requirement that the issue of creditor
connection affected whether a majority had voted for the CVA. Those creditors did not
stand in the same position as, and could not realistically be compared with, HMRC.
[18]
It was clear that the compromise of HMRC's claim (itself by far the largest of the
compromised claims) by the CVA had been achieved by the votes of creditors whose claims
were not themselves compromised and of those creditors who stood to gain a financial
advantage from completion of the broader Assets Sale Agreement, which required, as a
condition precedent, the compromise of claims including HMRC's by CVA or otherwise
at a level acceptable to the purchaser. Each of the creditor groups voting in favour of
the CVA derived advantages which were not available to HMRC. In Category B, a
compromised category, HSBC (the sole Category B creditor) would receive a substantially
higher payment in respect of a substantially smaller debt than HMRC on completion of
the transaction. In Category C, another compromised category, there were two creditors.
Moelis was by far the largest. It would achieve a significantly better recovery from the
overall transaction than HMRC. The other Category C creditor, Latham & Watkins, had a
claim of only £73,758,37, and its clients at least seemed to include a number of entities other
14
than PFML. In Category D, a compromised category, the creditors voting in favour were
other members of the group.
[19]
In categories F and G (non-compromised categories), the senior secured creditors
would not have the unsecured element of their claims, amounting to circa 94% of the whole,
compromised. By way of the intercreditor agreement, PFML would be sold and the secured
creditors would get the proceeds, in exchange for which they would discharge their debt
and their securities. The more a buyer paid, and the smaller any of the costs, the greater
the net proceeds available for distribution amongst the secured creditors would be.
Importantly, PFML was being sold free of debts. A buyer would pay far more for a
company without any debts than it would for a company with debts on the scale of PFML's.
The price achieved for the senior secured creditors was improved by this approach. A price
of $136m had been achievable only because the HMRC debt had been compromised. That
compromise was entirely to the gain of the secured creditors, who would receive sale
proceeds far in excess of what they would receive if the HMRC debt was intact. That was
a gain to those creditors which was achieved at HMRC's expense. It was achieved (and
indeed was only achievable) because the unsecured element of the secured creditors' claims
was so large that their votes drowned out HMRC's.
[20]
In Category I, consisting of non-compromised creditors, trade creditors were dealt
with. That was said to be necessary for the continuation of the business or to avoid
contagion, but little detail was given. It was unclear, for example, why it was said to be
necessary for Aberdeen City Council to be paid in full. It was said that failure to pay it
would have a "meaningful effect on the Council's ability to provide the public services
upon which the Company's operations... rely", but without similar reasoning applying to
HMRC's claim. The Category I creditors received a substantial advantage, both from the
15
non-compromise of their claims but also from the purchase of PFML and its continuation
in business through funding provided by a third party on the condition that HMRC's claim
should be compromised. It appeared to be suggested by the respondent that HMRC would
continue to receive payroll taxes as a result of the rescue of PFML. The larger part of these
taxes, however, was paid by employees, irrespective of the identity of the employer for a
particular contract, that employer's role being confined to accounting for the taxes to
HMRC. The Category I creditors' gain had been achieved at HMRC's expense.
[21]
It was also relevant to consider the nature of the HMRC debt that was compromised.
The debt arose from a historic failure to pay secondary NICs. That failure was the result of
an avoidance structure under which employees were transferred to a foreign employer for
the sole purpose of avoiding the relevant NIC liability. HMRC, unlike the other creditors,
was not a voluntary creditor in respect of that debt. It had been unable to choose the terms
upon which it transacted with PFML. Rather, PFML had accumulated a debt of very
considerable size through its own avoidance activity, which sum remained unpaid to HMRC
with the buyer of PFML being freed from its financial consequences as a result of the votes
of uncompromised creditors.
[22]
It was critical that a CVA was proposed in good faith and that the proposer should
be satisfied that it would not prejudice a creditor who was drowned out: Mourant at [88].
The treatment of HMRC's debt in the proposal failed to meet the standard of the greatest
care to ensure fairness which was required. The wider arrangements under contemplation
generated greater returns for the secured creditors, both because of the nature of those
arrangements, and because the secured creditors were creditors of PFML in respect of
guarantees giving them other rights of recourse. Even disregarding such other rights, the
proposed transaction would itself produce a recovery figure for the secured creditors (on the
16
premise that the sale proceeded) of 4.48%, as opposed to the 2.23% estimated for HMRC.
Those percentages fell to applied to very large sums of money. There was no obvious
commercial explanation as to why the secured creditors should receive, in relation to
the unsecured element of their debt, twice as much as HMRC. The explanation was that
they had managed to obtain the best recovery for themselves by using their majority to
compromise HMRC's debt. That was in their best interests, but there was no justification
for such unequal treatment of creditors.
[23]
In all the circumstances, the CVA was unfairly prejudicial to the interests of HMRC,
and the court should exercise its power under section 6(4) to revoke the arrangement.
Submissions for the respondent
[24]
On behalf of PFML, senior counsel submitted that if the CVA was revoked, then
PFML would enter into formal insolvency proceedings (ie, administration or liquidation),
in which its unsecured creditors (including HMRC) would receive virtually nothing.
[25]
The essential purpose of the CVA proposed by PFML was to facilitate a sale of
the Petrofac Asset Solutions division, including the share capital of PFML, to certain
buyers within the CB&I group of companies. The sale was agreed at the conclusion of a
competitive marketing process carried out by the court-appointed administrators of PFML's
ultimate parent company. As part of the sale, CB&I had required the dispute with HMRC
to be fully and finally resolved, so that PFML was released from the disputed claims. It
was easy to see why it took that position: the dispute had already dragged on for 14 years,
and the total value of the claim exceeded the entire cash consideration that CB&I would pay
to acquire the target assets. In addition, CB&I had required PFML's secured creditors to
release their claims, which amounted to more than £1 billion, and had required certain other
17
unsecured claims to be released. The cash proceeds of the sale would prima facie have been
payable to PFML's secured creditors, who held fixed and floating charges over its assets.
However, as part of the CVA, the secured creditors had agreed that a substantial part of the
cash proceeds of the transaction could be paid to the unsecured creditors whose claims were
compromised by the CVA. In terms of the CVA, those claims would be released in exchange
for a pro rata allocation of part of the cash proceeds of the transaction. Most of PFML's other
creditors by value, including its secured creditors, would also be required to release their
claims - not directly through the CVA, but as part of the wider transaction of which the CVA
formed part. The benefits of the CVA were derived entirely from the cash proceeds paid by
CB&I as part of the sale transaction, and from the secured creditors' agreement to allow the
creditors compromised by the CVA to receive part of the cash proceeds of the sale
transaction, to which they would not otherwise be entitled.
[26]
If the sale transaction did not take place, then PFML and the other operating
companies of the Petrofac Group in the Asset Solutions division would enter into
administration or liquidation. This would probably involve the immediate shutdown
of PFML's operations. That was not merely a theoretical risk: several entities within the
group had already entered insolvency proceedings and PFML was itself presently on the
verge of collapse, being forecast to run out of cash in the week commencing 16 March 2026.
The petition was issued by HMRC on the final day of the 28-day challenge period under
section 6 of the 1986 Act, thereby putting maximum pressure on PFML and allowing as little
time as possible to resolve the matter before it did run out of cash.
[27]
In the event of an administration or liquidation, virtually all of PFML's assets would
be distributed to its secured creditors, who held fixed and floating securities over them. It
was expected that the secured creditors would receive approximately $27,018,406 in a formal
18
insolvency of PFML. The unsecured creditors' returns would be confined to a share of the
"prescribed part" under section 176A of the Insolvency Act 1986 (amounting to £800,000 in
total, less the costs of distributing the same). This would generate returns for unsecured
creditors equal to only 0.12% of the face value of their claims.
[28]
The returns to the creditors whose claims were compromised by the CVA would
depend on the amount of the fund they were to share and the quantum of admitted claims
against that fund. Those claims would be far less than the total unsecured claims in a formal
insolvency process. There was an independent expert adjudication regime for resolving
disputed claims, including HMRC's claim. It was estimated that the fund would amount to
a total of £4,773,873, which would deliver a return of 2.23% on the claims compromised by
the CVA. That might prove to be a conservative estimate - but even on that basis, it was
18.5 times better than the expected returns on those claims in a formal insolvency scenario.
In any event, the fund would be no lower than £960,000,43, which would deliver a return
of 0.45% on the claims compromised by the CVA. That sum was greater than the amount
of the prescribed part in a formal insolvency process (£800,000), and there would also be a
material reduction in the total claims having to be satisfied. That was why the returns in the
worst possible scenario under the CVA (ie, 0.45%) were 3.75 times better than the expected
returns on the claims compromised by the CVA in a formal insolvency scenario (ie, 0.12%).
The net proceeds of the sale transaction would be remitted to the secured creditors. That
would deliver a return to them estimated to be approximately 4% of the value of their
outstanding claims. The balance of the secured creditors' claims against PFML (and against
the other target companies and assets being acquired) would be released. As a result, the
creditors whose claims were compromised by the CVA would obtain a return amounting
to greater than half of what the secured creditors stood to receive. That was a favourable
19
outcome for the unsecured creditors of an insolvent company with a huge amount of
secured debt, and one that was supported by the vast majority of PFML's creditors.
[29]
The CVA was approved by 85.9% in value of the creditors voting in person or by
proxy at the creditors' meeting on 30 January 2026. That was substantially in excess of
the requisite statutory majority of 75%. Indeed, the only reason that the CVA was rejected
by 14.1% in value of those voting was that HMRC was permitted to vote for the full amount
of its disputed claim. A total of 312 creditors, including 205 unconnected creditors, voted
in favour of the CVA. Only two voted against, with 18 abstentions. Aside from HMRC, the
only other dissenting creditor had a very small claim of £357,328, representing less than 0.1%
of PFML's debts, and had not challenged the CVA and or indicated any support for HMRC's
petition. Importantly, the CVA was approved by several unsecured creditors whose
undisputed claims were compromised by the CVA in precisely the same way as HMRC's
disputed claim, including HSBC, with an undisputed unsecured claim for circa £48 million,
and Moelis, with an undisputed unsecured claim for circa £7.4 million.
[30]
Section 6 of the 1986 Act empowered a creditor to challenge a CVA inter alia on the
ground that it unfairly prejudiced the interests of a creditor, member or contributory of the
company. To assess whether a CVA was unfairly prejudicial to the interests of a creditor,
two helpful tests (known as the "vertical comparison" and the "horizontal comparison")
had been developed by the courts: Lazari at [107] - [112] and the authorities cited therein.
[31]
In Re T&N Ltd at [82], David Richards J explained that in exercising its discretion
to sanction a scheme of arrangement under Part 26, where it was an alternative to a
winding-up, the court was unlikely to sanction a scheme:
"which was likely to result in creditors, or some of them, receiving less than they
would in a winding-up of the company, assuming that the return in a winding-up
would, in reality be achieved and within an acceptable time-scale."
20
The same idea appeared in the context of challenges to CVAs on the basis that they were
unfairly prejudicial, in the guise of the "vertical comparator" test: see Prudential Assurance
at [75] - [81] and Re AGPS Bondco plc [2024] EWCA Civ 24, [2024] Bus LR 745,
[2024] BCC 30 at [148] - [152]. In order to determine whether a vertical comparison revealed any
unfairness in the CVA, the court should simply compare the expected returns to HMRC,
in respect of the existing rights which the CVA sought to compromise, in the most likely
counterfactual scenario (ie, a formal insolvency process), against the likely returns to HMRC
in respect of those same rights under the terms of the proposed compromise contained in the
CVA, in return for which HMRC's existing rights were being extinguished. If the former
was greater than the latter, then no unfair prejudice could arise by way of a vertical
comparison.
[32]
HMRC did not appear to deny that it would receive a better return on its claim under
the CVA than it would receive in a formal insolvency process, nor that such a process would
be the inevitable consequence of the revocation of the CVA. Its argument that the difference
between the lowest outcome under the CVA (a return of 0.45%) and the likely outcome in an
insolvency scenario (a return of 0.12%) was minimal did not wash; as a matter of principle
the correct approach was to compare the likely outcome under the CVA to the likely
outcome under the insolvency scenario. On that approach, the CVA was likely to deliver
a return to HMRC (of 2.23%) that was approximately 18.5 times greater than the expected
recoveries in a formal insolvency process (0.12%). In any event, even on the worst possible
outcome under the CVA, it would still deliver a return to HMRC that was approximately
3.75 times greater than the expected recovery in a formal insolvency process. That fact alone
was sufficient to defeat HMRC's argument on the vertical comparator: Mourant at [67(d)];
21
Lazari at [108]. It was inaccurate to describe a 3.75x uplift as minimal. In any event, the
purpose of the vertical comparison was to identify the irreducible minimum below which
the CVA could not go, and the irreducible minimum was a return of 0.12%, which was
undoubtedly exceeded by the CVA.
[33]
So far as the horizontal comparison was concerned, the great majority of PFML's
creditors by value (including its secured creditors) would have their claims released by
the CVA or as part of the wider sale transaction of which it formed part. It was wrong to
suggest that the CVA was passed as a result of votes cast by creditors whose rights would be
unaffected. The opposite was true.
[34]
The horizontal comparison involved a comparison between the treatment of different
groups of creditors, and an assessment of whether any differences were justifiable. The
CVA regime was introduced to implement the recommendation of the Cork Report that
there should be a flexible regime that did not require pari passu treatment of all unsecured
creditors. Where a distressed or insolvent company was to be sold, the buyer would rarely
wish to take on all of the company's liabilities. Some of them might be impossible to pay,
and might indeed be the very reason why the company was distressed or insolvent. At the
same time, a buyer would not ordinarily wish to extinguish the liabilities owed to trade
creditors and suppliers. If that approach were to be taken, then the company might quickly
find itself without any suppliers (and thus without any business), thereby rendering it
worthless to the buyer and destroying the sale and the value it brought. For such obvious
practical reasons, it was very common for CVAs, schemes and restructuring plans to
exclude trade creditors and suppliers from the scope of the compromise: Re AGPS Bondco
plc at [170], Prudential Assurance at [88] and [90]; Mourant at [67(d)]; Debenhams at [103];
and Lazari at [111]-[112].
22
[35]
Some trade creditors might not be essential when viewed in isolation, but any
attempt to compromise their claims could cause other essential trade creditors to refuse
to supply the company. This "contagion risk" was a legitimate reason to avoid singling
out individual suppliers and trade creditors for separate treatment, even those that were
not essential when viewed in isolation: Debenhams at [107] - [108]. Furthermore, it might be
disproportionate or pointless to compromise a large number of relatively low-value claims.
In the final analysis, the question for the court was whether there was a reasonable
justification for the treatment of the various groups of creditors: Lazari at [111] - [112].
[36]
In the present case, the differential treatment of creditors was entirely justifiable.
HMRC's argument focussed on the treatment of the Category I creditors. They were to be
treated differently from the creditors whose claims were to be compromised by the CVA.
The Category I creditors would be kept whole. The question was whether the differential
treatment of the Category I creditors could be objectively justified. The great majority of the
Category I creditors (amounting to approximately £68 million or 76% in value of Category I)
were critical or irreplaceable suppliers. Compromising those claims would probably result
in the loss of such critical supplies, giving rise to contract breaches and the failure of the
business. The relevant suppliers included a substantial number of critical or irreplaceable
suppliers that were either engaged on current projects or likely to be needed for PFML to
deliver future projects, and there was no certainty that they could safely be replaced by
another contractor. A large part of PFML's business involved supporting offshore energy
assets. The relevant suppliers enabled PFML to operate safely and lawfully, such that any
compromise of their outstanding claims (leading to a loss of future supplies) would create
an immediate and unacceptable risk of operational failure or regulatory breach. More
generally, one of the defining characteristics of PFML's business was that it acted as a single
23
contractor to manage end-to-end delivery, centralising responsibility with PFML as main
contractor and removing the need for the client to manage multiple contractors directly.
As a result, the contractual and operational significance of the supply chain to PFML's
business could not be overstated. An inability to access the supply chain would result
directly in an inability to execute contracts and to provide integrated services to clients,
and would place PFML in breach of its existing contractual agreements. For those reasons,
it was considered that the sale transaction would not be viable if these Category I creditors
were to be compromised. Any attempt to compromise those creditors would cause severe
and immediate damage to critical supplier relationships at precisely the point when PFML
required supplier confidence to be maintained, in circumstances where CB&I's business
plan depended upon restoring and sustaining those relationships. That was an entirely
conventional justification for the exclusion of trade creditors and critical suppliers from the
compromise under CVAs, schemes or restructuring plans. HMRC was itself an essential
Category I creditor in respect of a separate non-tax claim for a breach of regulations made
under the Sanctions and Anti-Money Laundering Act 2018. That claim could lead to
criminal prosecution without a consensual settlement, and such a prosecution would
undermine the efforts to rescue PFML.
[37]
As to the second group within Category I, the contagion risk suppliers, that
amounted to only 6.7% by value of the claims in the category. PFML accepted that the
suppliers in this group could, in theory, be replaced. However, its management considered
that the compromise of such claims would give rise to a significant risk of "spooking" other
critical suppliers, leading them to withdraw their own supplies or tighten their terms to the
detriment of PFML's business. Aberdeen was a tight-knit, interconnected market in which
PFML was a substantial employer and where a substantial proportion of its suppliers were
24
based. Those suppliers had supported the business for two or more years during a
prolonged and difficult restructuring process. PFML had promised to pay them at the
end of the restructuring process, and that was the basis upon which they had continued to
supply the business. If PFML now attempted to compromise such suppliers, that would
have a chilling effect across PFML's entire group of suppliers, including those critical to
the business. The existence of a contagion risk was an entirely conventional justification
for excluding trade creditors and suppliers from the scope of a compromise, even if they
were not strictly essential to the business when viewed in isolation: Debenhams at [107].
[38]
The third group within Category I, de minimis amounts, represented only 4.5% of the
total claims in the category. Each claim within the group was less than £30,000. In light of
the total value of claims against PFML, such claims were regarded as de minimis, and the
cost of compromising the creditors in the group would be disproportionate to the benefit
obtained from doing so. The creditors in this group would have stood to receive dividends
of between £0.01 and £669 had they been compromised and given a right to participate in
the CVA as compromised creditors. PFML was notified by the CVA supervisors that, in
several cases, the cost of effecting such distributions would have been greater than the
amount to be distributed. There was nothing to be gained by compromising suppliers with
such small claims.
[39]
The fourth group within Category I, representing 9.6% of the total claims in the
category, were governed by Libyan law. They were excluded from the scope of the CVA
compromise because there was serious doubt as to whether a CVA could deliver a legally
effective compromise of debts governed by that law. PFML intended to compromise those
liabilities on a consensual basis instead. CVAs were normally used to compromise liabilities
governed by the law of England and Wales or of Scotland. A variation or discharge of
25
contractual rights in accordance with the governing law of the contract would usually
be given effect in other countries: Re Magyar Telecom BV [2013] EWHC 3800 (Ch),
[2014] BCC 448 at [15]. A Scottish CVA could deliver a legally effective compromise of English law
liabilities, since the CVA took effect under a UK Act of Parliament: see Phillips v Allan
108 ER 1120, (1828) 8 B&C 477, [1828] EngR 190;. In theory, a CVA could be used to compromise liabilities
governed by foreign law, at least if the relevant creditor voted in favour of the CVA, and
thereby agreed to be bound by it, or if the CVA would be recognised in the relevant foreign
jurisdiction. However, some jurisdictions (including Libya) did not make any clear
provision for the recognition of a Scottish CVA. In those circumstances, there was no point
in attempting to use a CVA to compromise the relevant liabilities.
[40]
As to the fifth group within Category I, those liabilities included local council
business rates and taxes payable to HMRC (in the form of VAT, PAYE and NICs) arising out
of PFML's continued trading operations. Such liabilities had been, and would continue to
be, paid in full as they fell due. The historical disputed liabilities owing to HMRC fell into
an entirely different category. They were 12 to 27 years old, did not arise out of continued
trading operations, and were disputed on legitimate and proper grounds. Most importantly
of all, CB&I required them to be compromised as a condition precedent of the completion of
the sale transaction. Given that that transaction was what would produce the entire benefit
of the CVA, a compromise of that claim was the only way to deliver those benefits.
[41]
Accordingly, there was a reasonable justification for excluding all five sub-groups of
Category I creditors from the scope of the CVA compromise. HMRC's challenge based on a
horizontal comparison had to fail.
[42]
The CVA was approved by several unsecured creditors with large claims that would
be compromised by it in the same way as HMRC's claim, including HSBC and Moelis.
26
Having regard to the votes cast by HSBC and Moelis, it was apparent that the CVA was
approved by reasonable creditors acting in good faith who would be compromised by it
in the same way as HMRC. That was significant: SISU Capital at [76], Prudential Assurance
at [96] supported the proposition that if a reasonable and honest man in the same position
as the challenger might reasonably have approved the CVA, the challenge would almost
certainly fail. The votes cast by creditors such as HSBC and Moelis were highly significant
and demonstrated that the basic test of fairness was satisfied.
[43]
HMRC claimed that its votes were drowned out by the very significant value of
the votes cast by creditors who were not compromised by the CVA, and by the votes of
compromised creditors who were connected companies. As a matter of principle, a CVA
could be passed by virtue of votes cast by creditors whose claims were not compromised
by it. That was an inherent feature of the statutory scheme and was not an automatic source
of unfair prejudice to the compromised creditors: Lazari at [156] and [186]. It was further
relevant to consider the extent to which the CVA was supported by creditors whose claims
would be compromised or released. One could imagine a hypothetical case where a CVA
was approved solely by those who stood to be paid in full, and the sole effect of the CVA
was to release the claim of an opposing creditor. Such a case would give rise to the potential
for unfair prejudice. However, the present case was fundamentally different. The CVA
was supported by a range of creditors who would be treated in the same way as HMRC
(eg HSBC and Moelis), and by other creditors whose claims would be released in parallel
with the CVA (ie, the secured creditors who would have their claims released as part of the
wider sale transaction of which the CVA formed part, receiving only an estimated 4% of
their whole claims). Only two creditors whose claims were to be compromised by the CVA
voted against it, namely HMRC itself and a small creditor with a claim of circa £357,000. A
27
total of 85.9% in value of all creditors voting on the CVA voted to approve it. In number,
312 creditors voted in favour of the CVA, and only two against. That indicated a very strong
show of support for the CVA and, objectively, that it was fair overall. The CVA was also
approved by the Category D creditors (ie, the compromised intra-group liabilities, which
would be treated in the same way as HMRC's disputed claim). Under the 2018 Rules, a
CVA could not be approved if more than half of the total value of the unconnected creditors
voted against it. In the present case, the CVA was supported by the vast majority of
unconnected creditors (approximately 83.27% by value of all unconnected creditors who
voted on the CVA). There was no basis for contending that the intra-group creditors did not
exercise their voting rights reasonably and in good faith. The relevant votes were cast by
separate companies in the Petrofac Group which did not have the same boards of directors
as PFML. They acted independently and at arm's length. Looking at the voting outcome
in the round, it was apparent that the CVA was strongly supported by PFML's creditors,
including those who were not connected with it, those who would be compromised by the
CVA, and those who would have their claims released in parallel with the CVA. HMRC
now stood alone in its opposition to the CVA.
[44]
HMRC finally claimed that its position was distinct from other creditors, in that it
was an involuntary creditor which had not been able to negotiate the terms on which it
transacted. However, as a matter of insolvency law, there was no principled difference
between an involuntary creditor and a voluntary creditor. Both were creditors with claims
eligible to be compromised, and neither was entitled to any special status in a CVA. Some
types of HMRC claims could rank as preferential debts in an insolvency, and such claims
were largely immunised by statute from the scope of a CVA compromise. The tax claim
in the present case was not a preferential debt, and the decision taken by Parliament was
28
to allow non-preferential tax claims to be compromised by a CVA in the same as all other
ordinary unsecured claims.
[45]
For all of those reasons, there was no proper foundation in law or in fact for HMRC's
contention that the CVA was unfairly prejudicial to its interests, and the court should refuse
to revoke it.
Decision
[46]
In an application such as the present, it is important not to lose sight of the basic
nature of the exercise being undertaken, namely a determination of whether or not the CVA
in question unfairly prejudices the interests of the challenging creditor. That core concept
of unfair prejudice, in this context as in others, requires the existence of prejudice (which
typically may arise from the terms of a CVA either when the result it produces for the
challenging creditor is not better or equivalent to the reasonably assessed value of the rights
being compromised, or when it yields to some interested parties, including the challenging
creditor, a return not at least equivalent to that yielded to others in a truly comparable or
less advantageous position) and that prejudice is unfair in some relevant sense.
[47]
Unfairness in this context falls to be assessed by asking whether, in all the relevant
circumstances, it is equitable to impose the bargain the CVA represents upon an unwilling
party. One way of approaching that question may be to ask whether an honest and
reasonable person in the position of the challenging creditor would hold out against the
imposition of the bargain, but that is simply one approach to the relevant question and may
not in itself furnish a complete or reliable answer in every case.
[48]
Relevant circumstances in assessing unfairness are likely to include, without
limitation, the financial backdrop to the CVA, the source of any bargaining surplus provided
29
by it, whether that surplus or its equivalent in value could have been made available in a
different way or in a different and more equitable pattern of distribution, and just how
homogenous or otherwise the rights of creditors with whom the challenging creditor invites
comparison may truly be.
[49]
I consider that having recourse to such open-ended standards, rather than to any
specific rule-based approach, is likely to provide a more accessible and reliable route to
the appropriate outcome than is reference to slightly abstruse concepts such as vertical
and horizontal comparisons, and may assist in avoiding the potential for sclerosis which
inherently attends the elevation of specific judicial observations apposite in particular cases
into statements of generally-applicable principle.
[50]
Turning to the case at hand, it appears clearly from the full and careful affidavits,
apparently furnished in all good faith, that the CVA was presented against a background
where there was, in effect, a moral certainty that the absence of a composition of PFML's
enormous debts would result in its imminent entry into a formal insolvency process, and
that there was no realistic prospect whatsoever of it being able to pay those debts or any
material portion of them by way of the profits of further trading. It further there appears
that the prospects of PFML being sold while still carrying its debts, or of a better overall
price being obtained for the assets to be acquired by CB&I, are negligible. Those features
of the case thoroughly permeate the issues to be considered, especially in that they provide
the background against which the value of HMRC's existing rights falls to be assessed, and
strongly suggest that any residual negotiating value which might otherwise attach to those
rights is properly to be regarded as inconsequential.
[51]
So viewed, there is indeed a bargaining surplus offered to HMRC by the terms of the
CVA. I accept the uncontroverted evidence for the respondent that the CVA will produce
30
for HMRC a return in excess of the reasonably-realisable value in insolvency of the rights
being compromised. That return will be at least 0.45% of the relevant debt, as opposed to
the 0.12% probably to be obtained in an insolvency process, and is likely to be more. While
I consider that it is the likely outcomes in both eventualities which fall to be compared, even
proceeding on the basis of a comparison between the likely insolvency outcome and the
minimum return from the CVA, those figures are not sufficiently close as to enable the
difference properly to be described as minimal, either in percentage or absolute pecuniary
terms. No prejudice to HMRC arises from this approach to the comparison between its
position under the imposed bargain and its position as that would otherwise be.
[52]
Turning to the question of whether the differential treatment afforded to different
categories of creditor of PFML under the CVA is unfairly prejudicial, it is certainly true that
the creditors in categories F, G, H and I do not have their claims compromised by the CVA,
and to that extent its terms may be considered prejudicial to HMRC. Creditors in the other
categories are, at least in point of form, treated in the same way as HMRC. It appeared to
be common ground that very little could be made of the treatment of the creditors in
Category H, since most of their claims would be compromised for nominal sums, albeit
outwith the ambit of the CVA itself, and - at least as a category - they would not leave the
larger transaction of which the CVA formed part with any greater return overall than
HMRC would obtain from the CVA itself.
[53]
The creditors in categories F and G are in a more nuanced position. Although their
claims are not compromised by the CVA (and their accompanying securities could not
have been released by it), the claims and the securities are to be cancelled in terms of the
inter-creditor agreement which forms part of the greater transaction including the CVA for a
return of 4.48% which, I was informed, was an estimate of what these creditors would have
31
received in respect of their securities in the event of PFML's insolvency. Is the prejudice to
HMRC constituted by that greater return to the secured creditors, outwith the terms of the
CVA, unfair?
[54]
The first thing to be observed is that these creditors are, by dint of their securities, not
in the same or a similar position to HMRC, an unsecured creditor. The second is that, given
that the immediate alternative to the CVA is an insolvency process, those securities would
be capable of being enforced to effectively the same result as the proposed return to these
creditors. A return to them which was less than that would provide no bargaining surplus
for them and, for reasons already given, would be unlikely to be forced on them by the court
even if the situation was such that the votes of other creditors could do so. If they were to
obtain that return by way of the insolvency route, HMRC would obtain a 0.12% return as
opposed to a minimum of 0.45% and a likely return of 2.23% to be obtained by way of the
CVA, resulting in a worse outcome for it. It follows that the return of 4.48% to the
Category F and G creditors cannot be criticised as unfair to HMRC. Any lesser recovery
than that for the Category F and G creditors would in all probability be capable of being
regarded as unfairly prejudicial to them.
[55]
In any event, given that the return to those creditors is not by way of the CVA, it
(and anything else that those creditors may receive in respect of their debts by way of other
rights of recourse) could only represent an unfairness to HMRC if the value it represents
might otherwise have formed part of the return proposed to unsecured creditors by the
CVA. Although it is reasonable to suppose that CB&I was prepared to acquire the assets
in question for a certain sum and would have been indifferent as to how that sum was
distributed amongst PFML's creditors, so long as the sums thereby owed to them were
all thereby extinguished, the Category F and G creditors, by dint of the sheer size of their
32
claims, their partially-secured nature and their other rights of recourse, would inevitably
have held the whip hand in the negotiations which led to the terms of the greater transaction
and the CVA being agreed. Without their willing participation, the transaction as a whole
could not have been unlocked. The affidavit evidence lodged on behalf of PFML is clear that
the deal that was negotiated, including the division of the funds made available by CB&I for
the discharge of debts within and without the ambit of the CVA, was the best available. It is
worth re-iterating that the 4.48% return to the Category F and G creditors is effectively what
they would have received in any event in an insolvency process. In these circumstances it is
difficult to see that their insistence on at least that level of recovery was a position from
which they could reasonably have been expected to resile, thereby making greater funds
available for distribution to others including HMRC. Any further recoveries which the
Category F and G creditors may obtain other than those deriving from the funds made
available from CB&I may be left out of account for present purposes, since there is no basis
upon which to suppose that they would ever have been made available to other creditors of
PFML.
[56]
There remains for consideration the position of the creditors in Category I, who are
to be paid in full as part of the arrangements in the CVA. I accept the explanations given
and recorded above for that course of action being taken as sufficient to indicate that that
treatment is not unfair to HMRC, with two quibbles. Firstly, I am sceptical that the concept
of contagion risk is of quite the universal application imagined by the respondent's
submissions. That said, the affidavit evidence provided on behalf of PFML indicated that
thought had been given to whether the contagion creditors being paid in Category I were
indeed likely to be influential in the fortunes of the company going forward, and no material
before me provided a basis upon which to discount the conclusion reached. Similarly, I was
33
not entirely convinced by the explanation tendered as to why PFML's business rates liability
to the local council was being paid in full when the historic liability to HMRC was not, but
in essence the answer is the same; the company has, in apparent good faith, reached the
conclusion that it is necessary or at least highly conducive to its expected ongoing trade
not to seek to negative that liability and no material which might serve to undermine that
conclusion was before me.
[57]
On the question of whether the votes of HMRC were "drowned out" by the votes of
creditors uncompromised by the CVA, it is certainly the case that the CVA was approved
by the votes of such creditors, primarily in Categories F and G, and to a lesser extent by
the votes of those in Categories H and I. It is, however, inherent in the nature of voting
weighted by the extent of debts owed that larger creditors have more influence in the
outcome than smaller ones. The process can only sensibly be described as "drowning out"
if some unfair advantage has been taken by the larger creditors of their position, which is
not the case for the reasons already stated.
[58]
I turn to express my views on some of the more minor issues raised in the parties'
arguments. Firstly, I do not accept that the attitude of HSBC or Moelis towards the CVA
can be taken as a surrogate for the attitude of an honest and reasonable creditor in the
position of HMRC. Unlike HMRC, they have recourse to rights of recovery against parties
other than PFML in respect of their debts, which, taken with their recovery under the CVA,
will produce for them an overall return of between 10% and 20% on the debts owed to them.
The attitude of such a creditor to the CVA cannot properly inform the answer to the
question of what attitude an honest and reasonable creditor without such other rights of
recourse would take to it. Although less was said to me about the exact position of the only
other compromised, unsecured and unconnected creditor voting for the CVA, Latham &
34
Watkins, that lack of information in itself prevents it being used as a relevant surrogate in
this context, and indeed no attempt was made on behalf of PFML to argue that it should
be. The same applies to the connected creditors who voted for the CVA. I consider that
HMRC's opposition to the CVA is certainly honest, and is a position that might be taken by
a reasonable creditor looking at the terms of the CVA in a relatively simplistic manner. That
I nonetheless do not accept that its challenge succeeds simply illustrates the limitations of
the "honest and reasonable" creditor test as a sure guide to the presence or absence of unfair
prejudice in this context.
[59]
Secondly, I accept that the fact that the relevant HMRC debt is disputed is something
that could in principle be taken into account in determining whether the treatment of that
debt in the CVA was fair or unfair, whatever the decision of the chairman of the relevant
meeting as to the value of the debt for voting purposes may have been. However, in the
present case the nature of the dispute, against the background of the tribunal decisions in
the apparently similar cases of other businesses in the same industry, is as matters stand
not such as to warrant any material discounting of the value of the debt for the purpose of
determining whether its treatment in the CVA was unfair or not. A knowledgeable and
prudent assessor of its value would not apply any such discount, and so this feature of the
case does not in the event carry any weight in the exercise which has to be carried out.
[60]
Similarly, I accept that the involuntary nature of the relevant debt is in principle
capable of being taken into account in assessing unfairness. However, any equitable force
that might otherwise inhere in that consideration is in my view very much attenuated by
the fact that the relevant debt was incurred in the period 1999 to 2014, but was not assessed
as due until 2020 and was, when that assessment was challenged, not even scheduled to
be determined by a tribunal until 2025. No explanation for that apparently very leisurely
35
progress in asserting and then enforcing the liability was provided to me. In those
circumstances it is reasonable to proceed on the basis that the matter could have been
determined long before now, and that the fact that HMRC is standing out of money in 2026
is referable at least as much to its own inaction in constituting and enforcing the debt as to
its original involuntary nature.
[61]
Finally, any suggestion that HMRC would stand to benefit from the rescue of PFML
by way of the receipt of future payroll taxes is too speculative to be given any weight in the
court's task of evaluating the fairness of the CVA. It remains to be seen whether PFML's
business will indeed be rescued in the medium to long term, and too many imponderables
attend the other possibilities in the situation to warrant serious consideration of the matter.
Conclusion
[62]
For the reasons stated, HMRC failed to demonstrate that the CVA unfairly
prejudiced its interests, and I refused the prayer of the petition accordingly.


BAILII: Copyright Policy | Disclaimers | Privacy Policy | Feedback | Donate to BAILII
URL: https://www.bailii.org/scot/cases/ScotCS/2026/2026csoh29.html