The Judicial Committee of the Privy Council Decisions

Contact Energy Ltd v. The Attorney General (New Zealand) [2005] UKPC 13 (23 March 2005)

ADVANCE COPY

Privy Council Appeal No. 14 of 2004

Contact Energy Limited Appellant

v.

The Attorney General Respondent

FROM

THE COURT OF APPEAL OF

NEW ZEALAND

---------------

JUDGMENT OF THE LORDS OF THE JUDICIAL

COMMITTEE OF THE PRIVY COUNCIL,

Delivered the 23rd March 2005

------------------

Present at the hearing:-

Lord Bingham of Cornhill

Lord Slynn of Hadley

Lord Hope of Craighead

Lord Rodger of Earlsferry

Lord Brown of Eaton-under-Heywood

[Delivered by Lord Rodger of Earlsferry]

------------------

1. On 5 July 1990 the Crown entered into a contract with the Electricity Corporation of New Zealand ("ECNZ") under which ECNZ agreed to purchase Maui Gas from the Crown. The commencement date of the contract was specified as 30 June 1989 and the contract was expected to run until 27 June 2009. In 1996 Contact Energy Ltd ("Contact") succeeded ECNZ as a party to the contract. Contact and the Crown are in dispute as to how the profit to which the Crown are entitled under the contract is to be calculated in the circumstances which have arisen. Both Goddard J in the High Court and the Court of Appeal (Gault P, Blanchard and Anderson JJ) held in favour of the Crown. Contact have appealed to their Lordships' Board. The hearing of the appeal was conspicuous for the clarity and economy of the submissions of counsel for both parties.
2. Under clause 9.3.1 of the contract Contact is required to pay the Crown a profit margin on the gas it purchases ("the Crown Margin"). The Crown Margin is to be fixed on each "Adjustment Date", which is defined in clause 1.2 as being the last day of each "Pricing Period". The pricing periods are defined in the same clause as meaning successive periods of 6 months each, the first being deemed to have begun on 1 October 1989. The Adjustment Dates - also referred to in the contract as "Adjustment Days" - are accordingly 1 April and 1 October. On those days the amount of the Crown Margin is calculated by a prescribed formula and is dependent on the quantities of gas actually taken and the quantities of gas specified in the contract. The definition of "Crown Margin" is set out in clause 1.2 of the contract:

"'Crown Margin' means, at any time, an amount for a GJ [a gigajoule] of gas equal to:

(a) NZD2.225 as adjusted by application of clause 9.5; less

(b) The aggregate of the Maui Gas Price and NZD0.45 [ie the ERL or Energy Resource Levy], in each case, for each GJ of gas."

3. Clause 9.5.1 contains the formula for adjusting the amount of the Crown Margin payable by Contact for each gigajoule of Base Gas over time:

"On each successive Adjustment Date the Crown Margin payable for each GJ of Base Gas under Clause 9.3.1 shall be adjusted by:

(a) multiplying an amount equal to NZD2.225 by the quotient obtained by dividing PPI by the Initial PPI; and then

(b) deducting from the product an amount equal to the Maui Gas Price and NZD0.45."

The terms "PPI" and "Initial PPI" are defined in clause 1.2. "PPI" means:

"on any Adjustment Day, the index number shown in the index for the last quarter in respect of which the index has been published and publicly released that ended no less than 3 months prior to such Adjustment Day: for this purpose the index is the Producers Price Index Inputs (all Industries) as from time to time published as table PPIQ.SI9 in the Key Statistics published by the Department of Statistics or the equivalent Government Department or other authorised agency required to prepare and publish such index or its equivalent and if that index is no longer published then such other index or its equivalent as may be agreed by and between ECNZ and the Crown and in default of agreement as to such index within 30 days of either party submitting an index to the other then such index as shall be nominated by the President for the time being of the New Zealand Society of Accountants upon the reference to him by either party."

"Initial PPI" is defined as meaning "1600, being the PPI for the quarter ended on 30 June 1989 (Base: the quarter ended 31 December 1982 = 1,000)".

4. The formula set out in words in clause 9.5.1 can be conveniently stated in this form:

NZD2.225 x PPI – NZD0.45 – Maui Gas Price = Crown Margin.

Initial PPI

The dispute concerns the first part of the equation only. It comprises two elements. First, there is the figure of NZD2.225, being the base price of the gas for escalation purposes (set at June 1989). Secondly, there is the factor, the PPI over the initial PPI, which is used to escalate the price as at the relevant Adjustment Date. The margin calculated in this way is then charged for the next six-month period.

5. As is apparent, the purpose of including the formula in Clause 9.5.1 was to maintain the June 1989 value of the Crown Margin by adjusting it by reference to a measurement of monetary inflation. At the outset of the contract the parties agreed to use the Producers Price Index Inputs (all Industries) as from time to time published as table PPIQ.SI9 ("SI9") in Key Statistics, published by the Department of Statistics or the equivalent Government Department. So long as SI9 continued to be produced and published in this way, no problem in calculating the relevant inflation factor could or did arise. But, when they entered into the contract, the parties foresaw that there might come a time when the relevant Government Department no longer published SI9. If that should happen, the parties agreed that they would use "such other index or its equivalent as may be agreed by" them and, in default of such agreement, then "such index as shall be nominated by the President for the time being of the New Zealand Society of Accountants upon the reference to him by either party".
6. In 1995 Statistics New Zealand ("Statistics") decided to redevelop the whole of the Producers Price Index. As part of this process, Statistics stopped publishing SI9 in Key Statistics after August 1996. None the less, Statistics continued to calculate SI9, and to provide it to people who asked for it, until the quarter ending 30 September 1999.
7. Instead of publishing SI9 in Key Statistics, from September 1996 Statistics published an interim index, called SAI9, pending the completion of the redevelopment. It was last published in December 1998 in respect of the quarter ending 31 March 1998.
8. In December 1997 Statistics announced that they intended to replace SAI9 with a fully redeveloped index called "SN9". It was released in August 1998, but was first published in Key Statistics in January 1999 to cover the quarters ending 30 June and 30 September 1998.
9. Statistics decided that, when they published SN9, they would include in it figures going back to December 1977, even though, of course, the new index had not been published at those earlier dates. For the period up until SI9 had stopped being published in Key Statistics the figures were to be taken from SI9. The interim index, SAI9 and, subsequently, SN9 were accordingly spliced on to SI9. For this purpose, the relative values of the figures in SI9 were not to be altered. But Statistics had chosen to adopt a base of 1000 as at December 1997 for SN9 and at that date the figure in SI9 was 1855. Therefore, in order to splice the two indexes together and to allow a seamless sequence of numbers to be published in the new table, Statistics scaled the figures in SI9 by a factor of 1000 over 1855. This had the result that, whereas the figure in SI9 for June 1989 was 1600, the equivalent figure for June 1989 in SN9 was 863. This particular change now gives rise to no dispute between the parties since Contact accept that it is not a requirement of the contract that the index used should have a base of 1000 at December 1982.
10. As already explained, when Statistics stopped publishing SI9 in Key Statistics, they none the less continued to calculate the Index and to make it available to those who wished to use it. It is now accepted that, even though SI9 was not published in Key Statistics, it was none the less being "published" for the purposes of the definition of PPI in clause 1.2 of the parties' contract. To begin with, SI9 was published alongside SAI9, but from January 1999 until September 1999, SI9 was published alongside SN9. In conformity with clause 1.2 of the contract, the parties continued to use the figures in SI9 for calculating the Crown Margin down to June 1999. The dispute between the parties concerns what was to happen when SI9 ceased publication after September 1999. The first and major dispute concerns the way in which the change from one index to the other was to be effected. There is also a secondary issue as to whether the relevant date for the changeover should be June or September 1999.
11. The parties began to negotiate on the question in April 2000 and, by 31 July 2000, Mr Peter Mackenzie for the Crown felt able to record that among the matters not in dispute was that the index to apply in the future would be "PPIQ.SN9 index as that index is described by Statistics New Zealand as the All Industries PPI Inputs index". On 10 August Mr Alex Love for Contact replied: "We confirm that PPIQ.SN9 is the applicable index". That measure of agreement was reflected in paragraph 3 of the Agreed Statement of Issues in the High Court: "After SI9 ceased to be published, the parties agreed to use SN9 for the purpose of calculating the Crown Margin as from and including the Adjustment Date of 1 April 2000".
12. Contact considered that, once the parties had identified SN9 as the appropriate index to use, only one thing had to be done. Clause 1.2 of the contract fixes the Initial PPI for June 1989 at 1600 but, due to SN9 having a base of 1000 at December 1997, the figure for June 1989 in SN9 is 863. Therefore, in order to be able to use the initial PPI of 1600, it would be necessary to scale the figures in SN9 by a factor of 1600 over 863 (= 1.854). That scaling would affect all the numbers in SN9, including those up to September 1999. But Contact accepted, of course, that, apart from the 1600 figure for June 1989, none of the other figures for that period would be used, since Contact could not reopen the relevant payments that had been calculated, in accordance with the contract, by reference to SI9. But, for December 1999 and subsequent dates, once scaled in this way, the figures produced by SN9 could simply be slotted into the formula for calculating the Crown Margin.
13. Discussions between the parties revealed that the Crown took a different view. As Contact saw it, the Crown did not wish to use SN9 in its published form, but, rather, a modified version of SN9 where the figure as at September 1999 (1014) had been increased so as to equate with the figure in SI9 at that date (1897). In other words, the figures given in SN9 for December 1999 and all subsequent quarters would be multiplied by a factor of 1897 over 1014 (= 1.871). Obviously, since Contact's approach resulted in the figures in SN9 being increased by a factor of 1.854, while the Crown's approach resulted in the figures from December 1999 onwards being increased by a factor of 1.871, adopting the Crown approach to calculating the PPI would give a higher Crown Margin not only for December 1999 but for all subsequent Adjustment Days. Contact objected that, in substance, the Crown were not intending to use SN9 but a composite index comprising SN9 for the past, but this modified version for all dates from December 1999 onwards.
14. On the basis that, despite appearances, the parties had therefore not actually agreed on the index to be used, Contact tried to refer the matter to the President of the New Zealand Society of Accountants in accordance with clause 1.2. And Contact still wish the matter to be referred to the President. As Mr Dobson explained in his submissions to the Board, Contact envisage that, while they would put forward SN9 as the appropriate index to use, the Crown might suggest that it would be better to use what Contact regarded as the modified version of SN9, linked at September 1999. It would be open to the President to nominate that modified index if he thought that it would be preferable in all the relevant circumstances. So, in the end, Contact might find that the modified index was specified for the contract. But that was a matter for the President to decide, since the parties had been unable to agree on the replacement index.
15. When Contact sought to refer the matter to the President, the Crown challenged his jurisdiction, on the ground that the index had in fact been agreed. They then raised the present proceedings for payment of $2,253,185.90, representing the difference between the amounts paid by Contact down to November 2001 and the amounts to which the Crown said that they were entitled on a proper interpretation and application of the contract.
16. Although the dispute between the parties presents itself as a dispute as to the appropriate new index to use or as to the way in which the transition to the new index is to be effected, that dispute only arises because of the difference in the way that SI9, on the one hand, and SAI9 and SN9, on the other, measured the relevant price levels. It appears that the difference arose because, in preparing the two indexes, Statistics used a different range of goods and services. In calculating SN9, they changed to a new range which they thought would better reflect the spread of goods and services which were then being bought and used. In that way it was hoped that SN9 would better reflect the prices which were actually affecting businesses and so would give a more accurate indication of the level of inflation. Since the range of goods and services being used to compile the two indexes was different, the results produced by the two indexes were also liable to be different. And Goddard J found that an unexpected divergence between SI9 and SAI9, which worked to the advantage of the Crown, did in fact occur in the March 1998 quarter. Once it had occurred, the higher figure for inflation was built into SI9 and continued up to September 1999. In terms of their contract, however, so long as SI9 was being published, the parties had to use it to obtain the PPI for the purpose of calculating the Crown Margin. If, however, when SI9 stopped being published, the parties simply moved over to use the figure in SN9 to give them the relevant PPI, in theory the result would have been either a net fall in the PPI by comparison with the PPI for the previous Adjustment Day, or else a lesser increase in the PPI than would have been produced if they had been able to go on using SI9.
17. The actual position was explained in the evidence of an expert statistician witness for the Crown, Dr P J Thomson. He said that if the parties simply switched from using SI9 as at September 1999 to using SN9 as at December 1999, this would produce a price escalation figure of 0.567% for that quarter, whereas SN9 itself showed that prices had escalated by 1.479% during that period. In his view simply switching from one index to the other in this way would produce a false price movement for that quarter. For this reason, merely substituting SN9 for SI9 would not give effect to the parties' intention that the price escalation up until September 1999 should be fixed by reference to SI9 and that the price escalation from September 1999 onwards should be fixed by reference to SN9. In effect, the parties would not be using SN9 (or indeed any other recognised index) to measure the price escalation in the September to December 1999 quarter. To avoid this, and to reflect the parties' intention that SN9 should be used to measure escalation from September 1999 onwards, it was necessary to link SN9 with SI9 in such a way that SN9 would be used to measure the escalation of prices in this quarter.
18. In his submissions for Contact, Mr Dobson did not dispute that substituting SN9 in the way that they proposed would have this effect. He argued, however, that for the purposes of the contract it did not matter whether the index used provided an accurate measurement of the movement of prices between any two quarters or Adjustment Days. The parties had not contracted on the basis that, say, a comparison of the PPI at June 1999 with the PPI at December 1999 would reflect the increase or decrease in the relevant price levels over that six-month period. They had contracted, rather, to use an index, SI9, to measure the increase in the relevant prices between June 1989 and June 1999 and, by virtue of the mechanism in the contract, they had agreed to use another index, SN9, to measure the increase in the relevant prices between June 1989 and December 1999. As required by the contract, Contact had paid, and did not seek to reopen, the Crown Margin for June 1999, which was produced by using SI9 to measure the increase in prices between June 1989 and that date. In accordance with the agreement between the parties, Contact had also paid the Crown Margin for December 1999, as produced by using SN9 to measure the increase in prices between June 1989 and this date. They had done, and would continue to do, the same for all subsequent Adjustment Days. Contact had therefore performed their obligations under the contract and nothing more could be required of them.
19. On this approach, the definition of PPI in the contract did not disclose an intention that the use of an index should produce a continuous linear sequence. Hence it was no objection to the simple substitution of SN9 that it did not maintain such a continuous linear sequence. In his reply for the Crown, Mr Kós accepted that the objective of the definition in the contract was not to produce such a linear sequence. In practice, however, that would usually be its effect even though exceptional blips might occur. For example, in the case of SI9, Statistics sometimes published a figure for one particular quarter which included a correction for an error that had crept into the calculation of the previous figures. In that situation the corrected figure would not precisely reflect the escalation in prices over that quarter – nor, in consequence, would the PPI. While he did not seek to support the Court of Appeal or Goddard J in so far as they had held that the contract mandated a continuous linear sequence, counsel nevertheless argued that, in entering the contract, the parties had chosen to use an index which could be expected in practice to produce such a sequence. Therefore, they would not have intended that the adventitious substitution of one index for another should be the occasion for disrupting this sequence.
20. If their Lordships had to decide the matter simply on the competing arguments about the function of an index in determining the PPI, they would see certain attractions in the approach favoured by the appellants. In particular, they see force in the point that the essential requirement is for the PPI to enable the parties to measure the increase in prices between June 1989 and the relevant Adjustment Day, even if, exceptionally, a comparison between the figures for that Adjustment Day and the preceding one does not precisely reflect the escalation in prices between those two dates. And, to the extent that Statistics may have considered that SN9 was a preferable way to measure the rise in prices between June 1989 and December 1999, there would be something to be said for simply using that more accurate measurement, even if there was a jump of some kind between the SI9 figure for September 1999 and the SN9 figure for December 1999. In their Lordships' view, such a result could not properly be characterised as arbitrary: it would be the combined result of the provisions of the contract and the work of Statistics in refining their figures.
21. The question does not fall to be answered in a vacuum, however. The contract does not contain anything which specifies how the transition from one index to another is to be effected. Moreover, the jurisdiction of the President of the Society of Accountants is strictly confined to nominating an index: he has no role in saying how it is to be applied. In that situation, their Lordships consider that, in default of any indication to the contrary, the parties to this commercial agreement may be supposed to have intended that, if they ever needed to change indexes, the change would be made in accordance with normal commercial practice.
22. One of the most striking features of the case is that, while Contact advanced subtle submissions on how the transition from SI9 to SN9 should be effected, they led no evidence as to the usual practice. By contrast, the Crown led expert evidence on the point, which Goddard J accepted. Faced with this major weakness in his case, Mr Dobson merely indicated that Contact had not been able to dispute that evidence. They could only say that it was open to the parties to agree to another course and not to follow the usual practice. That may be true, but there is nothing in the contract or in the evidence to show that they did. In paras 24 to 28 of her judgment, Goddard J outlines and quotes from the evidence of the Crown experts. Their Lordships therefore limit themselves to repeating the excerpt which she made, in para 28, from the evidence of Mr Ronald McKenzie, the Chief Economist, Business Statistics, for Statistics New Zealand. Dealing with the need to link any new index to the one that had been used before, he said:

"Not linking … would … cause dislocation – in the time period when the 'jump' is made from the old index to the new index, there would always be a 'jump up' or a 'jump down' in the new index. This jump up or jump down would not, however, be a price change reflected in any index; the old or the new.

In my 16 years' experience with Statistics NZ, I have spoken to or corresponded with numerous users who have contacted us for advice in the context of indexes being used in contractual escalation clauses. In that time, I have never known any users who have attempted a straight 'jump', or substitution from one index to another, upon a discontinuation of one of those indexes. The transition is always made through linking the two indexes together at a particular point in time" (emphasis added by Goddard J).

23. When she came to consider how the relevant terms of the contract fell to be interpreted, Goddard J had regard to this, and similar, expert evidence. Her Honour said, at para 72:

"I also find, on analysis of clause 1.2, that in expressly providing for the nominated contract index to be replaced, in the event that it is no longer published, clause 1.2 inherently provides for any replacement index to be 'linked' at the point at which the transition from the defined index to the replacement index is made. This finding is supported by the expert evidence, which unanimously establishes that linking is so integral to a transition from one index to another that it would be otiose to expressly spell out the need for it. As Mr Ronald McKenzie said: '… when moving from one index to another index the two indexes must by definition be linked' and 'a transition is always made through linking the two indexes together at a particular point in time'."

At para 55 of their judgment the Court of Appeal expressly accepted this finding. In those circumstances Mr Dobson did not argue that it was open to the Board to reject Her Honour's finding – nor indeed is there any counter-evidence on which he could have invited the Board to do so. Although the Crown witnesses were statisticians rather than commercial men, Mr McKenzie's particular expertise was in business statistics and he was able to speak from his long experience of dealing with businesses. In the absence of evidence to the contrary, his evidence was therefore a sound basis on which Goddard J could conclude that the need for linking between SI9 and SN9 would have been obvious to these commercial parties.

24. Any other conclusion involves accepting that the parties would have contemplated that one or other of them might benefit, not because of the need to correct an error in the official index which they had chosen, but simply because the Government Department happened to stop publishing that index and they were therefore forced to move to another index. From a commercial point of view, it appears to the Board to be a priori unlikely that the parties would have intended that an event over which neither of them had any control should have such an effect. Their Lordships are therefore not surprised to find that the expert evidence, as to the way businessmen always link one index with another, confirms that the commercial approach would indeed be to smooth the transition and so prevent either party profiting from the need to change indexes.
25. Whether the matter is expressed in terms of clause 1.2 inherently providing for any replacement index to be linked to the original index at the point of transition, or in terms of an implied term to the same effect, appears to the Board not to be of critical importance in this case. Mr Dobson pointed out that the particular implied term envisaged by the Court of Appeal was defective since it was limited to an index being linked to SI9, whereas it would have to apply whenever two indexes had to be linked. That may be so, but it is equally clear that an implied term could be formulated so as to apply generally.
26. For these reasons their Lordships would reject Contact's main ground of appeal.
27. Mr Dobson submitted that, even if it was necessary, for the purposes of the contract, to link SI9 and SN9 in the manner envisaged by the Crown, the courts below had been wrong to hold that the link was to be made at September rather than June 1999. The dates are significant because the figure in SI9 for September was higher than for June and linking the indexes at September produces a higher PPI. According to the definition in clause 1.2, the PPI means the index number shown in the index for the last quarter in respect of which the index has been published not less than 3 months prior to the relevant Adjustment Day. Since the Adjustment Days are 1 October and 1 April, the index figure for the quarter ending 30 September of any year can never be used for the PPI. So the last index number in SI9 which would have been used by the parties to establish the PPI would have been the number for June 1999. Therefore, if the indexes were to be linked, counsel argued, it should be at June, not September 1999.
28. Their Lordships reject this argument. It is plain from the definition of PPI in clause 1.2 that SI9 remained the relevant index for the purposes of the contract for so long as it was published. The parties are now agreed that, in terms of clause 1.2, SI9 was last published in September 1999. It follows that, under the contract, the transition to SN9 could not take place before 30 September 1999. To link SI9 and SN9 at any earlier date, such as 30 June 1999, would therefore contradict an express term of the contract. To link them at 30 December 1999 would be too late, since SI9 ceased to be the specified index at 30 September. The appropriate date is, therefore, 30 September when SI9 was superseded by SN9. Of course, this means that the link is made at the date of the publication of an index figure which could never have been used to provide a PPI. That does not matter, however, since the linking of the indexes is a separate matter from their use. In principle, therefore, there is nothing to prevent the indexes being linked at a date when the original index would not have been used to provide the PPI.
29. For these reasons their Lordships will humbly advise Her Majesty that Contact's appeal should be dismissed with costs.