Guardian Assurance Limited v The Commissioners for HMRC

This hearing was conducted remotely by video. Prior notice of the hearing had been published on the gov.uk website, with information about how representatives of the media or members of the public could apply to join the hearing remotely in order to observe the proceedings. As such, the hearing was held in public.

In this case the Appellant, Guardian Assurance Limited (“GAL”), appeals against a closure notice issued by the Respondents, HMRC, on 30th August 2018 in respect of an enquiry into GAL’s company tax return for the accounting period ended 31st December 2013, which amended GAL’s tax returns for the periods ended 31 December 2013, 31 December 2014 and 31 December 2015.

The closure notice increased GAL’s profits subject to corporation tax by bringing into account dividend income of £2.7m and an increase in the value of a shareholding by £96.4m. In addition, the consequential amendment made by HMRC (a) to the return for the period ended 31 December 2014, HMRC brought into charge additional dividend income of £3.4m and share value increases of £19m; and (b) the return for the period ended 31 December 2015, HMRC brought into account a loss on disposal of £63.9m and foreign exchange hedging gain of £17.8m.

The only issue in this appeal is whether, during the relevant period, GAL’s shareholding in a Canadian company called Empire Life Financial Services Limited (“ELFS”) (the “Shareholding”) was a “structural asset” and, therefore, “long-term business fixed capital” within the meaning of section 137 of the Finance Act 2012 (“FA 2012”).

The Appellant’s skeleton argument foreshadowed a dispute around the Respondents statement of case and whether this was required to be amended. As it turned out and by the time this issue came to be ventilated, the Appellant, sensibly, abandoned any objections to the way that the Respondents had set out their case in the statement of case and/or the way that it was now being pursued.

In this hearing I had the benefit of witness statements produced by Mr. Jonathan Yates, a former CEO and executive director of GAL, Mr. Andy Pennington, GAL’s former With Profits Actuary, Chief Actuary and Actuarial Function Holder, and Mr. Matthew Taylor, an accountant and specialist advisor in the insurance sector, for the Appellant. All three witnesses gave oral evidence and were tendered for cross-examination (albeit the opportunity was not taken up in respect of Mr. Taylor).

In addition to the witnesses of fact I heard expert evidence from Mr. Dumbreck, a fellow of the Institute and Faculty of Actuaries, for the Appellant and Mr. Turnbull, also a fellow of the Institute and Faculty of Actuaries, for the Respondents. Both expert witnesses gave oral evidence and were tendered for cross examination.

Naturally, there was an opportunity for me to ask questions of any witness.

I was also taken to a significant number of documents contained within the hearing bundle to which I also had regard.

In the UK insurers typically conduct their business through companies limited by shares (also referred to as “proprietary companies”) or mutuals. Proprietary Companies are owned by, and run for the benefit of, their shareholders, just like any other company. Mutuals have no shareholders and are owned by, and run for the benefit of, their current and future members, who are some or all of the policyholders.

Business undertaken by insurance companies is usually categorised as long-term insurance business (life and pension business, critical illness cover and permanent health insurance) and general insurance (accident and health, property, liability, marine aviation and transport, credit and suretyship). Generally, an insurer will write either long term or general business, but not both.

When an insurance company enters into a life insurance contract it undertakes, in return for a premium, to pay out a lump sum or regular payments on death or diagnosis of a critical illness in respect of the person whose life forms the subject matter of the policy.

Life policies can either be “with-profits” (also known as “participating”) or “non-profit”.

Non profit (“NP”) policies refer, usually, to policies where the benefits to the policyholder are fixed according to the terms of the insurance contract and do not depend on the exercise of insurer discretion.

Under a “with-profit” (“WP”) policy all amounts receivable from WP policyholders are paid into a segregated WP fund of the life insurer and then invested. The policyholder is usually entitled to at least a guaranteed amount on maturity of the policy (the guaranteed sum). In addition, profits made on investments of the WP fund may be paid to policyholders in the form of discretionary bonuses. These bonuses may take the form of annual bonuses (which once declared cannot be taken away and form part of the guaranteed amount) and final or terminal bonusses which are added on maturity to the policy to top up the guaranteed amount.

After assessing all the evidence, and acknowledging that the Appellant carries the burden of proof, I make the following findings of fact upon the balance of probabilities.

GAL was incorporated in 1893 and since that time has been a principal life insurance company in the Guardian corporate group.

In 1980 GAL acquired a 74% shareholding in Montreal Life Insurance Company (“MLI”) from Guardian Royal Exchange Assurance Ltd, GAL’s ultimate parent company. By 1983 this shareholding had increased to 99% of the ordinary share capital and also included some preference shares.

At least by 1988 GAL had a life insurance business which consisted almost entirely of WP business conducted through its WP fund and the WP fund held the shares in MLI. The WP fund was a “90:10” fund, which meant that policyholders were entitled to 90% of the surplus from the fund and the shareholders were entitled to 10%.

In 1986 GAL entered into a joint venture arrangement in Canada with E-L Financial Services Corporation (“ELFC”). At the time ELFC owned two Canadian insurance companies: Empire Life Insurance Company and The Dominion of Canada General Insurance Company. ELFC and GAL agreed to consolidate MLI’s life insurance business with that of Empire Life Insurance Company under a newly incorporated, unquoted Canadian joint venture holding company called E-L Financial Services Ltd (“ELFS”). Originally GAL owned 21% of the share capital of ELFS and 79% was held by ELFC. Subsequently, ELFS acquired the life insurance division of The Dominion of Canada General Insurance Company in exchange for further shares issued in ELFS. This increased ELFC’s holding in ELFS to 81% and diluted GALs holding down to 19%.

GALs relationship with ELFC was governed by a shareholders agreement entered into by the parties on 30 June 1986 (the “ELFS shareholders agreement”). The ELFS shareholder agreement provided, inter alia, that:

The initial board should consist of 17 members (which number could be increased at any time by the board) and so long as GAL continued to own at least 10% of the outstanding common shares in ILFS it would be entitled to nominate four members to the board.

GAL was entitled to nominate two out of six members of the board’s Joint Advisory Committee.

GAL was entitled to access company records, including the ability to require monthly financial reports.

The ELFS shareholding was, in the first instance, acquired and held by GAL’s WP fund.

In 1988 the GAL WP fund was closed to new business and has been in “run-off” since.

The evidence shows that for most of the period between 1986-2008 GAL nominated one or more member to sit on the board of ELFS. The extent to which these nominees attended board meetings or participated in decision making / discussion is entirely unclear.

On 1 May 2008, GAL appointed Mr Paul Weiss as a non-executive director of Empire Life/ ELFS. He continued to serve in this capacity until, at least, around 2012. He attended some board meetings and provided some reports to GAL in this capacity. During this time his reports lacked detail and he only provided very limited (high level) information to GAL.

In November 2011, following the acquisition of the Guardian Group by Cinven, a private equity firm, GAL carried out a restructuring of the WP fund. This involved, with the consent of the Prudential Regulation Unit and the Conduct of Business Unit of the FSA, GAL establishing a new NP fund. The ELFS shareholding was transferred for £32m together with a right for the WP fund to further consideration if ELFS should be sold for a higher value within five years. The approach to pricing was subject to review both by Deloitte LLP and Ernst & Young.

Following the acquisition by Cinven, it is fair to say, GAL became more proactive in the management of ELFS. Jonathon Yates and Andrew Birrell joined the board of ELFS- becoming honorary directors in June 2012 and full voting members of the board in April 2013. Mr Yates and Mr Birrell, thereafter, regularly attended board meetings between 2012 and 2015 and [were involved in the strategic management of ELFS]. They] had input into the capital and risk management as well as the dividend policy. It is likely that this enabled a larger dividend, than might otherwise have been the case, to be paid to GAL.

In GALs 2013 financial statements it increased the carrying value of the ELFS shareholding, following a revaluation, to £128.4m. It then further increased this to £147m in its 2014 financial statements. The Appellant says that this resulted from the higher dividends and the improvements made to ELFS’s capital and risk management. The Respondents say that the revaluation occurred because the Appellant discovered Embedded Value Reports which valued GALs shareholding in ELFS at £155.92m in 2010 and £161.98m in 2011. In my judgment it is likely that the revaluation resulted from a combination of ‘discovering the value’ (particularly by discovering the Embedded Value Reports) and ‘improving the value’ (particularly by increasing the dividends). It is fair to say that the latter is likely to have had a much less significant impact.

GAL completed its company tax return for the period ended 31 December 2013 on the basis that the ELFS holding was ‘long-term business fixed capital’ within the meaning of section 137 Finance Act 2012.

The financial statements of GAL for the period ended 31 December 2014 valued the ELFS holding at £147 million. GAL completed its company tax return for the period ended 31 December 2014 on the basis that the ELFS holding was ‘long-term business fixed capital’ within the meaning of section 137 Finance Act 2012.

In December 2015, GAL sold the ELFS holding to ELFC for c. £96 million.

GAL completed its company tax return for the period ended 31 December 2015 on the basis that the Shareholding was ‘long-term business fixed capital’ within the meaning of section 137 Finance Act 2012.

In January 2016 Cinven sold GAL to Swiss Re Group.

In the end, I did not find it necessary or helpful to consider the expert evidence that was led in this hearing and I have, therefore, decided against burdening this judgment with a summary of it.

Before turning to the relevant provisions and the way that they fall to be interpreted, it is necessary to set out, in a little detail, the backdrop to the relevant legislation.

Prior to 1 January 2013, the regime for the taxation of long-term insurance business was contained within Part II of the Finance Act 1989 and Part XII ICTA 1988 (plus sundry other provisions). It is common ground between the parties that, during this period, trading profits of life insurance companies were calculated on the basis of regulatory returns made to the Financial Services Authority, rather than on the basis of statutory accounts.

It is further common ground that “Life companies” were taxed on the “Income minus Expenses” (I minus E) basis, which effectively charged profits made by the shareholders and the investment return arising for the benefit of certain policyholders to tax. Three categories of insurance business were recognised for tax purposes, and a life insurance company's investment income, gains and losses were allocated between those categories on the basis of statutory rules. However, increases and decreases in the value of “structural assets” held by insurance companies in a non-profit fund were not taken into account as receipts and expenses in computing trade profits.

The definition of structural assets was given at section 83XA(3) Finance Act 1989 (now repealed), which provided that:

The taxation of life insurance companies was substantially reformed by FA 2012 in anticipation of fundamental changes to the regulatory reporting framework by the EU Solvency II Directive. With effect for accounting periods beginning on or after 1 January 2013, the provisions of Part 2 Finance Act 2012 replaced the regime for the taxation of long-term insurance business within Part II Finance Act 1989 and Part XII ICTA 1988 (see section 148 Finance Act 2012).

Under the new rules, trading profits would be calculated on the basis of life companies' statutory accounts, rather than on the basis of regulatory returns (as had been the case under the former regime principally contained in Part II Finance Act 1989 and Part XII ICTA 1988).

Section 113 Finance Act 2012 provides that:

This ensures that any income generated by a life assurance business from the “long-term business fixed capital” is excluded from the computation of the trade profits of its life assurance business, together with any increase or decrease in the value of such capital.

Section 137 Finance Act 2012 (which apply to assets held on or acquired after 1st January 2013) provides as follows:

Section 63 Finance Act 2012, which is headed, “Meaning of “long-term business” and “PHI business” provides that:

As set out at the outset of this judgment the only real issue in this case is whether the ELFS shareholding constituted a “structural asset” and, therefore, “long-term business fixed capital” within the meaning of section 137 of the FA 2012.

The proper starting point of any discussion must, in my view, be to define what is meant by “structural asset” in the present context. The term is not further defined in the legislation and has not, apparently, received any previous judicial scrutiny. I must, therefore, necessarily look elsewhere for assistance.

No assistance, in my judgment, can be derived in the present case from looking at the dictionary definition of “structural”. The Oxford English Dictionary defines the word as “a component or material having a structural or load-bearing role ,esp in a building”. In the first instance the definition is almost entirely circular. Secondly, the term seems to generally be used in the context of buildings (of whatever form or ‘structure’). The fact that the OED does not assist is hardly surprising given that, in my view, the term “structural asset” as used in section 137 of the FA 2012 is a technical term of art. It is used in the context of the taxation of some highly specialised circumstances and it is somewhat surprising that the term is not further defined.

Neither is any real assistance to be found from the commentary provided by Mr. Taylor in his witness statement which largely concerns itself with the consultation process which led to the FA 2012. There are clear and obvious dangers in placing weight upon a statement from an individual involved in the consultation process when attempting to interpret legislation. The idea, I think, is that this will assist in divining Parliament’s underlying purpose and intention in enacting the relevant legislation. However, a person involved in the consultation process seems to me to be too far removed from Parliament to be able to offer any useful input on its purpose in enacting the legislation. Divining the purpose of legislation is a hazardous task at the best times and is made more difficult, not easier, by the introduction of witnesses such as Mr. Taylor. The use of witnesses to aid interpretation of legislation seems to me, in general, to be an unwelcome development.

However, I agree with Mr. Peacock that assistance may be derived from the predecessor legislation to section 137 of the FA 2012 (namely section 83XA Finance Act 1989 as amended by the Finance Act 2007). In particular, the explanatory notes to the Finance Bill 2007 explain as follows:

At the Committee Stage for Finance Bill 2007, the relevant Government minister said:

It is argued by the Appellant that the relevant minister used the term “structural assets” in distinction to “assets expected to be turned over in the course of a trade” and that this latter reference is a reference to circulating capital (or assets of a revenue nature) (see John Smith and Sons v Moore [1921] 2 A.C. 13 pp 19-20 ). Therefore, it is argued, fixed assets which are not part of the circulating capital used to meet current liabilities, which are held for the long-term, and matched to policy liabilities are structural assets. I must disagree. One needs to read the entirety of the sentence. The minister begins by saying that structural assets are “assets which are held by an insurance company as part of its trading structure”. This sentence encapsulates the core of what is meant by “structural asset”. It is then contrasted with assets that are “turned over”. This illustrates and emphasises the fact that such assets (ie those liable to be turned over) can never, by their nature, be structural assets. It does not, however, suggest that the relevant legislation should be read to conclude that all assets of a fixed capital nature held for the long term are “structural assets”. Neither is Mr. Peacock assisted in this argument by the fact that for the purposes of s.137 the definition “long term business fixed capital” is broken down into two constitutes parts such that the asset is (a) held for the purposes of its long term business and (b) is a structural asset. It does not follow that because s.137(a) repeats the term “long term business” that the second requirement contained s.137(b) is synonymous with “fixed asset”. If Parliament had intended to define “long term business fixed capital” as an asset (a) held for the purposes of its long term business, and (b) comprising fixed capital, then it would have said so. The fact that it chose not to do counts against rather than for the Appellant.

In summary, it seems to me to be clear that a “structural asset” means an asset held as part of the relevant insurance company’s trading structure. As such it is almost inevitable that such an asset will comprise fixed as opposed to circulating capital, but not all fixed (as opposed to circulating) capital assets will be “structural” assets. Accordingly, any analysis of capital vs revenue expenditure will simply add an unnecessary layer of complexity and would serve to distract from the focus on the actual wording of section 137.

First, the Respondents argue that the ELFS shareholding was “employed and at risk in the GAL business” and that it is well-established in case-law that where an asset is acquired for the purposes of an insurance company’s business, is employed in and is at risk in that business, the investment income should be treated as trading income and taken into account in computing trading profits. This might well be the case in the present case (although I make no finding about this) and, therefore, the income derived from such an asset might well fall to be treated as trading income. However, whether the income derived from the ELFS shareholding is trading income is to ask and answer a slightly different question. Parliament has chosen to require that consideration be given to whether or not the asset in question is a “structural asset”. In my judgment focussing on this question alone is likely to yield the right answer. The authorities referred to by the Respondents in support of its contention ( Liverpool and London and Globe Insurance Co v Bennett , Owen (HM Inspector of Taxes) v Sassoon (32 TC 101), Punjab Co-operative Bank Ltd., Amritsar v. Commissioner of Income Tax, Lahore [1940] & General Reinsurance Co Ltd v Tomlinson (Inspector of Taxes) Alherma Investments v Same [1970] 1 WLR 566) are, therefore, simply not in point.

Secondly, the Respondents argue that the fact that the policyholders of GAL’s long-term business benefited disproportionately from the ELFS shareholding is a strong indicator that the shareholding was not ‘structural’ to its business. This argument is predicated on the basis that the ELFS shareholding was held at the relevant time in the WP fund such that 90% of the dividends or profits went to the policyholders and only 10% went to the shareholders. It is not clear to me how an analysis of the benefits to be derived from holding the asset, as between shareholders and policyholders, informs as to the nature of the asset viz a viz the business (and in particular as to whether or not it is a structural asset). For my part I do not find such an analysis helpful.

Thirdly, the Respondents contend that the ELFS shareholding was acquired pursuant to an intended change on the part of GAL from controlling and managing an operating entity in Canada to holding an investment. In support of this argument the Respondents seek to draw a distinction between an investment and what is said to be a structural component of the business. The argument is, I think, based upon the perceived passive nature of investments and I will deal with this particular point later in this judgment. For the moment, I can see no justification in imputing an intention on the part of GAL to change from controlling and managing an operating entity to holding an investment. The evidence suggests an intention on the part of GAL to enter into a joint venture (albeit as the more junior partner) and this is apparent from, at the very least, the ability of GAL to nominate members to the board of ELFS. It does not assist the Respondents to argue that GAL negotiated terms allowing for easy exit from the venture. That is to naturally be expected from a junior partner to the venture.

Fourthly, the Respondents point to the fact that GAL held insufficient economic and voting rights to have any control over the strategic or operational activities of ELFS. At all material times GAL held only a minority shareholding in ELFS. Further, at all material times, the ELFS shareholding was treated as a financial investment in GAL’s audited financial statements, rather than as an associate. The significance of this is that International Accounting Standard 28 (2011) (“IAS 28”) provides that if the investor holds less than 20% of the voting power of the investee, it is presumed that the investor does not have significant influence, unless such influence can be clearly demonstrated. That ELFS was not, at all relevant times, an associate (for the purposes IAS 28) appears to be accepted by GAL. Control or significant influence over an asset seems to me to be an important factor when one considers whether the asset is structural in relation to the business. One would, generally, expect a business to exercise control and influence over an asset that was structural to its business. In the present case, whilst only having a minority holding, GAL was able to exercise some (more than minimal) control and influence over ELFS through its ability, in particular, to appoint board members. The fact that it did not effectively exercise this control or influence for part, at least, of the time goes to its ability to manage the asset rather than to the nature of the asset itself. Equally, care needs to be taken in transposing conclusions under IAS 28 to a consideration of the whether the asset is structural to the business. It is axiomatic that the two tests are different. In any event, this is only one factor that falls to be considered in the mix and limits on such control and influence would not necessarily be fatal to the conclusion that the asset was a “structural asset”.

Fifthly, the Respondents submit that, read fairly, the evidence makes plain that for the overwhelming majority of the period of its ownership of ELFS GAL had little or no involvement in the ELFS business and limited information regarding the nature of its activities, which is inconsistent with the asset having formed part of the structure of the GAL business. In my view it is, indeed, fair to say that the evidence suggests that for the period [May 2008 – June 2012] GAL had little involvement or information relating to the ELFS business. Mr. Weis’s role (as the director nominated to the board of ELFS by GAL) was largely ineffectual (a point acknowledged by Mr. Yates during cross-examination) because he provided very limited information to GAL and was left largely to his own devices. That position improved significantly with the appointment of Mr. Yates and Mr. Berrill as honorary and then full board members. The evidence shows that they were much more involved and unearthed important information. Although much time was spent at the evidence stage of this hearing on his issue, it seems to me that, once again, involvement in the management of the asset or possession of information about it is a factor (very much allied and akin to the exercise of control and influence) that goes into the mix when one considers whether the asset is a structural asset or not. In some sense, the control and influence on the one hand and involvement and information on the other are joined at the hip. One would expect a degree of involvement and information in order to exercise influence and control. The distinction between an “investment” asset and a “structural” asset is not, in my judgment, helpful. The point that is really, it seems, being made is in relation to the passive nature of most (but not all) investments; but that point is already covered under involvement with the management and information about the asset.

Sixthly, the Respondents content that (a) the tax treatment adopted by the Appellant, for periods prior to 31 December 2012, in relation to the ELFS holding is informative of GAL’s views as to the function of the holding in its business, and (b) to infer that the catalysts for the change in tax treatment adopted in the Appellant’s tax return prepared for the period ended 31 December 2013 were the increased valuations of the ELFS holding in or about 2012 and 2013. Neither point is a good one. The tax treatment adopted by the Appellant for periods prior to 31 December 2012 is irrelevant not only because the law was different then, but also because it does not tell us anything about the classification of ELFS holding (i.e. whether it is a “structural asset” or not). Likewise, whether or not one infers that the catalyst for the change in tax treatment was the increase value attributable to the ELFS holding does not go to answering the core question – namely whether the ELFS holding was a “structural asset”.

Lastly, the Respondents argue that the circumstances surrounding the sale of the ELFS shareholding demonstrate that, contrary to the impression given by the Appellant that the sale was a matter of ‘last resort’ (as would be likely be the case for the sale of an asset that is structural to the business), the sale was an opportunistic attempt to realise value from the ELFS shareholding, which is inconsistent with the ELFS shareholding having constituted a structural asset of the GAL business. This criticism, to my mind, is unfounded. The evidence does not demonstrate a prior intention on the part of GAL to sell ELFS. Indeed, the evidence shows that it is more likely that Cinven wished to retain and eventually sell GAL as an entire unit (including the ELFS holding). The sale of GALs ELFS holding to ELFC might well be described as opportunistic. However, this does not mean that ELFS was not a structural asset for GAL and clearly Swiss Re Group’s views on whether or not the ELFS shareholding was structural to the business going forward are irrelevant to the consideration.

In summary, there are only really two factors which point to the ELFS shareholding not being a ‘structural asset’ employed in GALs business at the relevant time. The first is the limited control or influence that GAL could exercise on the ELFS business by reason of its minority shareholding and the second, allied, factor is limited involvement in, and information about, the ELFS business. However, these factors are not hard edged. Rather, there is a sliding scale where at the one end we have complete control, influence, involvement and information and at the other end no control, influence, involvement or information. In addition, as I have said, neither factors, either individually or taken together, are determinative of the issue.

There are two further important factors of particular relevance in the present case. First, it seems to me that where we are dealing with an asset consisting of a shareholding in another company then it will be important to examine the nature of the business being carried on by the investee / subsidiary company. If it is the sort of business that is the same or similar or which in someway operationally enhances the business of the relevant company then that is a factor that points to the holding being structural to the business of the relevant company. In the present case ELFS was itself a life insurance company, and as such, the business of ELFS can very much be seen as an adjunct to the business of GAL.

Second, the length of ownership or holding of the asset seems to me to be an important factor. The idea that an asset has assumed structural importance in a business connotes, in my view, a sense of permanence. Therefore, the longer that an asset is held the more likely it is that it will have become a “structural asset”. However, like all the other factors identified above it is but one factor which must be considered. The ELFS Shareholding was held by GAL for almost 30 years (from 1986 until 2015).

Taking into account all of the above factors in the round and weighing them up in the mix, I come to the conclusion that the ELFS shareholding was a “structural asset” for GAL at all material times.

For the reasons given above I allow the appeal in full.

This document contains full findings of fact and reasons for the decision. Any party dissatisfied with this decision has a right to apply for permission to appeal against it pursuant to Rule 39 of the Tribunal Procedure (First-tier Tribunal) (Tax Chamber) Rules 2009. The application must be received by this Tribunal not later than 56 days after this decision is sent to that party. The parties are referred to “Guidance to accompany a Decision from the First-tier Tribunal (Tax Chamber)” which accompanies and forms part of this decision notice.