The Legal Representative of Succession Paul de Maroussem v Director General, Mauritius Revenue Authority (Mauritius)

This appeal is concerned with the taxation of a gain realised by the sale in 1988 of 75 arpents of development land at Wolmar on the west coast of Mauritius. The land was part of an estate of about 1,000 arpents (other parts of which had already been sold for development). The area of 75 arpents included a disused sandpit and its infilling was an important element in various infrastructure works undertaken before the morcellement (subdivision) of the area into 456 housing plots. These were sold to members of the public, the prices depending largely upon proximity to the sea. On each sale the purchase price was payable by a down-payment followed by a payment of instalments, with interest at 14% per annum. The instalments were meant to be paid in full within three years but in practice some payments were still being made eight years later.

Three parties were involved in the development: the developer, Société Roger de Chazal (“the Société”), which undertook responsibility for the infrastructure works and the general management of the project; Medine Sugar Estate Co Ltd (“Medine”) which owned the estate, but subject to a 99-year lease of which 50 years were unexpired in 1988; and Mr Paul de Maroussem (“the taxpayer”) who was the owner of this lease. Under written agreements entered into between them in December 1988 the Société was to bear all the costs of the project and receive one half of the proceeds of the sales (after land transfer tax). Medine and the taxpayer were to share the other half in proportions of 49.96% and 50.04% respectively. These were the proportions in which compensation had been divided between them under an arbitration on the compulsory purchase of another part of the estate earlier in 1988.

This appeal is therefore concerned with a sequence of events which began over 20 years ago. The litigation has been protracted for a number of reasons. Initially the taxpayer did not recognise that he was or might be under a tax liability arising out of the sales. In June 1995 an assessment for 1991-92 was made on him under the Income Tax Act 1974 (“the 1974 Act”), and in June 1997 further assessments were made on him for all the years from 1989-90 to 1994-95 except the year which had been already assessed. The later assessments were made under the Income Tax Act 1995 (“the 1995 Act”), which repealed and replaced the 1974 Act with effect from 1 July 1996 (in Mauritius the tax year starts on 1 July and ends on 30 June).

The taxpayer appealed against all these assessments, but unfortunately he died shortly before the hearing of these appeals. The appeals were taken over by his heirs (who are included where appropriate in references to the taxpayer). The taxpayer failed in his appeals to the Tax Appeal Tribunal and in further appeals to the Supreme Court, but was successful in his final appeal to the Board, whose opinion ([2004] UKPC 43, [2004] MR 213) was delivered by Lord Scott of Foscote on 22 July 2004. The matter was remitted to the Supreme Court to be disposed of in accordance with the Board’s opinion. The sequence of events down to 2004 is fully set out in paras 1 to 30 of the Board’s opinion, to which reference may be made for further detail.

As Lord Scott explains in the opinion, Mauritius has no tax on capital gains as such. But section 11(1) of the 1974 Act required some particular receipts to be brought into account as “gross income” (that is, income before allowable deductions), including in para (h):

Had the Board acceded to either party’s extreme position disposal of the appeal (in relation to the valid assessments) would have been straightforward. Had the MRA won, the assessments would have been confirmed; had the taxpayer won, the matter would have been remitted to the Tax Appeals Tribunal to perform the ministerial exercise of amending the assessments to exclude any receipt from the sales of development land. The decision that the Board actually reached meant that further steps which were not merely ministerial (and might prove contentious) would be necessary. In the event they have been so contentious that the litigation is now before the Board for a second time.

At the end of its opinion the Board set out its conclusions (paras 46 and 47):

The Mauritius Revenue Authority Act 2004 made important administrative changes in the tax system. It established the MRA under a Director-General. The Tax Appeals Tribunal was replaced by the Assessment Review Committee (“ARC”) whose procedure was regulated by Part IV of the Act (which came into force on 1 July 2006) and by the Assessment Review Committee (Appeal) Rules 2007. Section 19 provided for written representations to be lodged by the taxpayer with ARC; section 20 provided for the hearing of representations by ARC; and section 21 provided for an appeal from a decision of ARC by way of case stated to the Supreme Court.

It is to be noted that the Board’s conclusions set out in para 7 above were expressed in tentative terms, and the Board drew attention to the fact that they had not had any submissions from counsel as to the appropriate procedure. The pending changes in the tax system may have contributed to the uncertainty. When the matter was mentioned to the Supreme Court on 10 October 2005 (as appears from a transcript of the short hearing) Mr Basset SC referred to a letter dated 11 August 2005 from the Chief State Attorney to the taxpayer’s attorney. That letter is not in the record but it seems to have put forward figures for chargeable income and tax due for the outstanding years of assessment. Senior state counsel moved for an order directing the Commissioner of Income Tax to issue a revised assessment for the outstanding years, adding,

Mr Basset said that he had no objection to that, though he had a complaint about the tax paid for the years for which the assessments were out of time. Neither side referred the Court to any provision of the 1995 Act or of the Tax Appeal Tribunal Act 1984. Matadeen J (who was sitting with Domah J) is recorded as having disposed of the matter as follows:

He then repeated the disposal in slightly different words:

Very shortly afterwards, on 13 October 2005, an official acting for the Commissioner wrote to the taxpayer as follows:

On 8 November 2005 the taxpayer joined issue with the Commissioner by lodging representations with ARC. The grounds were set out in three paragraphs of the notice, all of which challenged the Commissioner’s estimate of the profit element, based as it was on the government valuer’s valuation of Rs33,750,000 as the 1988 value. This was described as a gross under-valuation unsupported by any evidence. The taxpayer put forward a competing figure, Rs72,147,545, as the 1988 value.

The matter came before ARC at a series of hearings on 7 February, 14 March, 14 May, 21 May, 28 May and 6 June 2007. Both sides were represented by counsel. Oral evidence was given for the taxpayer by a valuer, Mr Dilmohamed, and by Mr Mayer, an experienced property developer who was associated with the Société. He gave evidence that the venture had cost just over Rs34m in infrastructure and other costs, the largest item being about Rs26.65m for the earth-moving contractors. The Director-General called the government valuer, Mr Ramlagan. The evidence of the valuers is considered below. It was common ground that the project was a great success. All 456 lots were sold in a single weekend. The Commissioner’s detailed schedule shows that the capital receipts amounted to about Rs147.4m, and the total interest to just under Rs30m.

ARC gave a decision in favour of the Director-General. Its written ruling concluded,

The main issue before ARC and before the Supreme Court was the issue of valuation, but some procedural points about the system of tax assessment have appeared intermittently in the litigation, including one entirely new point (on section 132 of the 1995 Act) which the taxpayer sought to raise for the first time before the Board. It might be sufficient to say that the new point has been raised far too late (having already been conceded), and that Mr Basset did not ultimately press the other points (on which his argument involved resiling from a position on which the taxpayer had succeeded, in relation to the two out-of-time assessments, before the Board in 2004). But since the two points are of some general importance to the tax system in Mauritius it seems better to deal with them briefly.

Section 132 of the 1995 Act provides as follows:

But that is not the correct analysis. Amendment of an assessment under section 132 (normally to increase the amount of the tax charged) is something that would be expected to happen before the assessment has been made the subject of an appeal to ARC (or before that body existed, the Tax Appeal Tribunal). At a hearing ARC itself has jurisdiction (under section 6(4) of the Tax Appeal Tribunal Act 1984) to “confirm, amend or cancel any decision” (including an assessment) made by the Commissioner. If at an appeal hearing ARC amends an assessment under section 6(4) (say, by reducing an assessment from Rs500,000 to Rs400,000) it is not acting under section 132 but in exercise of its own adjudicative function.

The Board’s decision in 2004 established that the Tribunal had erred in law (in short, by treating section 11(1)(h) of the 1974 Act as taxing the whole of the taxpayer’s receipts, rather than his gain, from the development). The assessments therefore needed to be adjusted downwards, but there was at that stage no factual evidence on which to determine the amount of the downwards adjustment. So it was necessary for the matter to go back to ARC unless the parties could agree the figures; and initially the ball was in the Commissioner’s court to decide how large an adjustment he was prepared to accept. His letter of 13 October 2005 and its accompanying schedules stated his position, and led with very little delay to the taxpayer’s application by way of representations to ARC. But there was no formal amendment of the assessments under section 132, nor was there any need for any amendment. The assessments under appeal remain those made in 1995 and 1997, but the Commissioner had made clear the limited extent to which he now sought to uphold them, and the basis on which he sought to do so. It is true that the Supreme Court referred to this as “reassessment”, but there is no need to construe this as meaning a formal assessment for which there was no need, and which the Commissioner had no statutory authority.

The point which Mr Basset rightly does not press is whether section 130 of the 1995 Act enables an assessment to be made (subject to the four preceding years limit) in respect of a year before the 1995 Act came into force on 1 July 1996. This point was assumed by the Board in 2004 (para 23 of Lord Scott’s opinion) and the assumption was in favour of the taxpayer, as the previous time limit (under section 79 of the 1974 Act) had applied only to the amendment of assessments more than six years back.

The argument the other way was that section 162 of the 1995 Act repealed the 1974 Act as from 1 July 1996, and that section 163 provided for the 1995 Act to come into operation –

The Board considers those cases to have been correctly decided. The coming into force of a taxing statute involves questions of substance (what transactions are to be taxed, and at what rates?) and questions of procedure (what are the procedures and time limits for assessing tax?) The presumption against a statute having retrospective operation applies to the former but not in general to the latter (subject always to the language of the particular statute) and in the Société Bahemia & Co case the Supreme Court followed that principle in construing and applying sections 129, 130, 162 and 163 of the 1995 Act.

That may not quite dispose of the procedural issues, since the only substantive question raised by the taxpayer with ARC was valuation, and Mr Basset made plain that he objected to an ‘assessing penalty’ included in the Commissioner’s computations, and he may have other points to raise on those computations. This should not be taken as any encouragement to further expensive litigation; the litigation has gone on too long already. But if any outstanding points cannot be agreed they should be referred to ARC promptly in a form (what is the correct assessment on the taxpayer for each outstanding year?) which produces finality, and not merely a further step towards finality.

The valuers for both sides were seeking to identify the market value of the land, including the market’s view of its development potential, as at 1988 immediately prior to the commencement of the venture. The higher that value was identified as being, the lower would be the figure for the tax payable. As already indicated, the valuations produced were starkly different, Mr. Dilmohamed for the taxpayer putting the value at Rs.72,147,545, and the government valuer, Mr. Ramlagan, putting forward a figure of Rs.33,750,000. The ARC had to resolve this issue, which turned principally upon a difference in the methods used by the two valuers.

The government valuer adopted the approach of looking at “comparable” transactions, seeing what value per arpent each transaction produced and adjusting so as to allow for differences in the characteristics of the land in question and the date of the transaction. He referred to three comparables, all concerning land with potential for residential development and all in the same area of Mauritius, namely Flic en Flac. One was a sale of 54 arpents of land on 27 February 1987 at a price of Rs.74,074 per arpent. The second was a sale of 96 arpents of land on 12 September 1988 at Rs.203,125 per arpent, and the third was an award of compensation for the compulsory acquisition of 49 arpents as at 3 September 1986. That threw up a figure of Rs.360,000 per arpent. Making various adjustments, the government valuer arrived at a figure for the subject land of Rs.450,000 per arpent.

Mr. Dilmohamed for the taxpayer used a different valuation method, which he described as the residual method. Guided by “the actual amount received” for the sale of the subject land in plots (see the ARC case stated), namely Rs.147,615,542, he took a figure of Rs.147 million as the gross realization from the sale. From this he deducted the actual cost of infrastructure, Land Transfer Tax and developer’s profit. No allowance was made for any risk factor as all the plots were in fact sold in one weekend. This gave a net value for the land of Rs.72,147,545, equivalent to a figure of just under Rs.962,000 per arpent. He had, he told the ARC, the benefit of hindsight, although he acknowledged that a purchaser in 1988 would not have had that knowledge. Nonetheless, he criticised the government valuer’s use of the comparables method, because the subject land was a unique site.

The ARC, as its case stated shows, rejected the evidence of Mr. Dilmohamed and preferred that of the government valuer, stating that the former was wrong to have used actual figures for gross receipts and for costs. That did not necessarily reflect what a potential buyer cum developer would have expected to achieve in 1988, as opposed to what was actually achieved subsequently. The ARC accepted the use of the comparables method by the government valuer and accepted his valuation, though it appears to have considered it somewhat generous towards the taxpayer, given that all the comparables produced a lower figure per arpent.

The taxpayer’s appeal to the Supreme Court raised a number of valuation points, but in essence it was contended that in the circumstances the residual method of valuation was the only correct method and that the government valuer had made use of the wrong comparables. The Supreme Court rejected these arguments. It held that the government valuer had been entitled to use the comparable method of valuation and that the comparables were appropriate in characteristics and timing, subject to making the necessary adjustments. Relying upon this Board’s opinion in Mon Tresor and Mon Desert Ltd v Ministry of Housing and Lands [2008] UKPC 31, [2008] 38 EG 140, it concluded that the residual method should be reserved for exceptional cases and that in any event the taxpayer’s valuer had wrongly applied such a method, in particular by using actual figures “when the method implies a hypothetical development.” Those conclusions are now challenged by the taxpayer.

The appeal to this Board comes, like the appeal by way of case stated to the Supreme Court, only on a point of law. It is important to be clear as to the approach to be adopted in the different stages of the process which may arise under the tax system in Mauritius. The Director-General, formerly the Commissioner, is required by section 129 of the 1995 Act to make assessments “according to the best of his judgment.” If the matter goes to the ARC, that body acts as the tribunal of fact within the framework of law. It proceeds to find facts de novo and can simply find that the assessment in question is incorrect, whether or not there is any legal flaw in it. But if an appeal is to succeed before the Supreme Court, what is required is an error of law on the part of the ARC.

It is said on behalf of the taxpayer in the present case that the ARC erred in law by accepting the “comparables” method of valuation as opposed to the residual method. Both are, of course, methods of valuing land used by expert valuers. In both methods, the valuer is seeking to arrive at the value of the land in question on a particular basis, in the present case the open market value of the subject land in 1988 immediately before the morcellement scheme. That open market value would reflect such development value as the land was seen by the market at that time as possessing. The choice of method by which a valuer arrives at such a market value does not automatically or indeed even normally give rise to an issue of law. The Board would emphasise the need for caution, lest differences in methods used by opposing valuers are inappropriately elevated into issues of law.

It is true that issues of law can sometimes arise if the valuer (or any subsequent valuation tribunal) has made some assumption contrary to the legal basis upon which the valuation is to be made or has erred in one of the ways to be found set out most conveniently in the well-known Wednesbury case, even though that was a case on the proper exercise of discretion: Associated Provincial Picture Houses Ltd v Wednesbury Corporation [1948] KB 223 at 229. Thus if a valuation body has left out of account a relevant consideration or has taken into account an irrelevant one, or has arrived at an irrational valuation, a point of law may well arise, but the test of irrationality creates a formidable hurdle. It is surmounted only if the valuation is one at which no reasonable valuation body could have arrived if directing itself in accordance with the law.

Mr. Basset for the taxpayer accepted that challenge. He put his main argument on valuation in terms of irrationality and submitted that the comparables relied on by the government valuer and the adjustments made to the values they produced were such that irrationality was established. He pointed out that one of the comparables relied on threw up a value per arpent that had to be multiplied six-fold to get to the value used by the government valuer. Another had to be doubled. The government valuer gave no adequate explanation for how he adjusted these figures. It followed that no reliable basis for the use of the comparable method existed. Consequently, it was argued for the taxpayer, the residual method should have been accepted, as proposed by his own valuer. A subsidiary point was raised in oral argument, to the effect that the ARC, like the government valuer, took into account the existence of a sitting tenant of the subject land, namely the taxpayer, and this was in the circumstances an irrelevant factor as the respondent accepted.

The Board is not persuaded that any case of irrationality has been made out. The use of comparables is the normal method of valuing land, as the Board recognised in the Mon Tresor case. It described the method in the following terms, at para 7(b):

It went on to add, at para 7(e):

That remains the view of this Board. Of course, a comparable can rarely be used without making some adjustments to allow for differences between the transaction used as a comparable and the subject land. Location, size, ease of development, potential development and date of the transaction may all be among the factors requiring some adjustment to the values thrown up by the comparables. That in itself is not a reason for rejecting the comparable method. In the present case the government valuer had used comparable transactions from close to the date required for valuing the subject land. Those transactions were of large areas of land, ranging from 49 arpents to 96 arpents, and the sites were all in the Flic en Flac part of the island. Those characteristics would seem to render those transactions appropriate as comparables.

The main criticism advanced on behalf of the taxpayer was that those transactions threw up significantly lower values per arpent than the government valuer had used, an unusual argument since the logical conclusion from it could well be that his value should have been lower and the resulting tax assessment higher. But in any event the Board does not see that as a valid criticism. It is not right simply to take a single comparable and to use that to assert that the value per arpent was a sixth of the eventual value per arpent placed on the subject land. The valuer’s task is to look at all the relevant comparable evidence available to him and then to arrive at a judgment based on all that evidence and his own professional experience. Here he had a range of transactions available to him, including one which produced a figure of Rs.360,000 per arpent, a figure still below that of Rs.450,000 which he eventually used in his valuation but not one very far below. The Board can see no basis for regarding the “comparables” method of valuation as inappropriate, since proper comparables were available, nor does it regard the eventual valuation using that method as in any way irrational.

As for the residual method of valuation, one of the main reasons why that has been regarded generally as less appropriate than the “comparables” method is that it usually requires even more speculation about future events than the latter method. If one is to use the residual method to ascertain what a willing vendor and a willing purchaser would have agreed as the price of a piece of land at a given date, assuming that the land has development potential, one has to make an estimate of many things. One needs an estimate of gross receipts after sale or development, plus some estimate of the timing of those receipts. One needs estimates of the cost of the project, whether it be a project of sale or of development, together with an estimated allowance for developer’s profit and for any tax payable on the project itself. Then the residual figure may indicate the value of the land in the open market prior to the start of the project. But it is readily apparent that such an exercise is fraught with uncertainties. Hence the method is one which has tended to be used when the use of the land being valued is one for which open market comparables are simply not available, such as a public use for which there is no open market. That was not the situation in the present case, where comparable transactions did exist, albeit requiring a range of adjustments.

As already indicated, the taxpayer’s valuer sought to overcome the inherent uncertainties in the residual method by making use of the actual receipts and costs of the project available to him at the later date. As he conceded in evidence, he employed hindsight. The Board agrees with the Supreme Court that that is an erroneous way of applying the residual method. The hypothetical vendor and purchaser in 1988 before the start of the morcellement scheme would have known none of that information used by the taxpayer’s valuer. They would have had to estimate or guess, with unknown degrees of accuracy, the receipts and costs. They cannot have attributed to them the knowledge of subsequent events, such as the sale prices fetched, the speed of sale of the parcels, the costs of the scheme and so on. The fact that the taxpayer’s valuer resorted to this improper application of the residual method simply emphasises the difficulties inherent in that method and the absence of any irrationality in the ARC’s decision to reject it.

That deals with the taxpayer’s main points on valuation. As indicated earlier, an argument was raised orally before the Board to the effect that the ARC was wrong to take account of the fact that there was a sitting tenant on the land, namely the taxpayer, who should have been seen as someone who would willingly surrender his interest in order to benefit from the scheme. This does not seem to have been a matter explored before the ARC as the fact-finding tribunal. The government valuer was not cross-examined about it, nor was it a complaint ventilated in the eight grounds of appeal to the Supreme Court. Consequently it was not an issue with which the Supreme Court dealt. It is too late to raise such a criticism for the first time before this Board. That conclusion does not involve any injustice to the taxpayer, since there is force in the ARC’s observation that, in the light of the comparables, the government valuer’s valuation was, if anything, generous towards the taxpayer.

In the Board’s opinion there is no error of law to be found in the decision of the Supreme Court. It follows that the appeal must be dismissed with costs.