Shun Fung Ironworks Ltd v. Director of Buildings and Lands (Volume Ii)
Read the full judgment text of on BabelCite..
1. The issue here is the annual volume of rebars SFI might reasonably have been expected to make on the basis of there not having been the threat and fact of resumption. We consider that question in relation to the No-Scheme-World alternatives of SFI's likely tonnage of good product on the basis of its plant and machinery as installed in 1982, and its likely tonnage on the basis of likely improvements to its plant and machinery.
|
Crown Lands Reference No. 18/87
IN THE HONG KONG LANDS TRIBUNAL ----------------- Between
----------------- VOLUME II ----------------- PART 2 PARTICULAR ISSUES
SECTION I : SFI's REBAR-MAKING CAPACITY 1. The issue here is the annual volume of rebars SFI might reasonably have been expected to make on the basis of there not having been the threat and fact of resumption. We consider that question in relation to the No-Scheme-World alternatives of SFI's likely tonnage of good product on the basis of its plant and machinery as installed in 1982, and its likely tonnage on the basis of likely improvements to its plant and machinery. 2. There is agreement on the approach to be adopted. 3. Basically, one needs to work out, in days, the amount of time the works was likely to be out of production in the course of a year on account of stoppages resulting from such factors as holidays, breakdowns of equipment, waiting time, planned maintenance, and interruptions in the flow of materials during the production process due to poor coordination between departments. 4. According to the government, based on the 1982 plant and machinery installation, SFI could reasonably have expected to lose the equivalent of 127.95 days per year from such causes, leaving only 237.05 days for production (a.k.a. "operating days"). SFI says it would have lost only the equivalent of 89.03 days, leaving 275.97 days for production. 5. Having determined the number of operating days available for production, one then has to calculate the number of heats per day on the basis of the electric arc furnaces' melt-down times for making 22 M/T of liquid steel. 6. Although theoretically the EAFs had a nominal capacity of up to 25 M/ts of liquid steel per heat, we do not think that, in practice, SFI, in the No-Scheme-World would have got beyond melts of twenty-two tons of liquid steel before upgrading the EAF's by adding water-cooled panels - a development which, we find, later in this section, would have occurred by 1985/86. 7. We are aware that the liquid steel capacity of the EAF's could have been increased by enlarging their internal diameter through the use of shorter refractory bricks for the lining, but we do not think it more likely than not SFI would have done that in the No-Scheme-World. 8. The government has contended that the 22 M/T of liquid steel is the sole constant of any significance amongst the various factors to be applied in our calculations. With that, we do not agree. The amount of scrap to produce 22 M/T of liquid steel is also a constant factor, the amount of such scrap being 1.086 times the liquid steel, with the result that 23.91 M/T of scrap is required to produce 22 M/T of liquid steel. 9. There is agreement between the parties on the formula to be applied for working out each electric arc furnace's melt-down time. 10. EAF 1, which has a 8,250 KVA transformer, requires 425 KWh/ton of scrap charge for melt-down (to 1550oC), whereas the corresponding figure for EAF 3, with its 12,000 KVA transformer, is 410 KWh. (At this point we note that, throughout much of the hearing, both sides proceeded on the assumption that EAF1's transformer was 8,500 KVA and all the calculations in the case were done on that footing. Rather than set the accountants' and steel experts' computers a-whirring yet again to work through all the consequential amendments, we have left the calculations as they are on the practical ground that the difference in the award, were the recalculation to be done, would be trivial).
11. Taking into account the load cycle, the continuous KVA during meltdown can be increased by 20%. 12. With charge tons of scrap taken as 1.086 times the liquid steel (based on a 92% yield), the calculation is as follows: -
13. The methodology is Mr Medley's, from his report 30/01 page 30. 14. In the earlier stages of the hearing, when both sides were working on the basis of SFI's yield of liquid steel from scrap being only 90% (instead of the 92% subsequently agreed), the workings of the formula assumed the scrap charge tons were 1.11 times the liquid steel, which meant that 24.44 M/T of scrap was needed to produce 22 M/T of liquid steel. On that basis, the melt-down period for EAF 1 was 1.61 hours, and, for EAF 3, it was 1.10 hours. 15. Based on the now-agreed yield of 92% for liquid steel from scrap, EAF 1 achieves melt-out 108 seconds (1.61 hrs - 1.58 hrs) faster than when the assumed yield was 90%; EAF 3 is 72 seconds faster: (1.10 hrs - 1.08 hrs). 16. The quicker the melt-down, the more heats per day, and, hence, the higher the production capacity. 17. Somewhat surprisingly, the government strained at this particular gnat, despite having swallowed several of what we regarded as rather large camels, and did not concede the quicker melt-down time and its consequences. 18. In our view, the quicker melt-down times follow as a matter of inexorable logic from the agreement that the scrap to liquid steel yield was 92%, rather than the 90% assumed earlier. 19. Hence, we find as a fact, that the melt-out time for EAF 1 was 1.58 hrs, and for EAF 3, 1.08 hrs. 20. After melt-down, the next step is the refining and superheating of the liquid steel. There is agreement that occupies 0.33 hrs (i.e. 20 minutes). 21. The next element to consider is "power-off time". There are two types of "power-off time", one being "useful", and the other, "non-useful". 22. "Useful" power-off time covers the essential processes, with the electricity for the EAF turned off, which have to be carried out in relation to the electric arc furnace, either invariably, or at least frequently, during its heat cycle, commencing with the charging of the first basket of scrap into the electric arc furnace, and finishing with the fettling, minor repair of the EAF's refractory bricks, and electrode adjustment or changing which follow the tapping of the moulten steel from the EAF into a ladle for casting. 23. "Non-useful" power-off time occurs when the electricity to the EAF is turned off during a scrap - melting cycle, because of some non-regular happening, such as, for example, an EAF breaking down. 24. Such "non-useful" power-off time falls within the availability or utilisation factors, some of which can apply to the works as a whole and others to various sections of it. we will elaborate on availability/utilisation factors in due course. We have already mentioned how, according to SFI's reckoning, 89.03 days a year are lost for production purposes on account of such matters as equipment breakdowns, and interruptions in the flow of materials. The government's equivalent figure is 127.95 lost days. "Non-useful" power-off time is treated as falling within those lost days. 25. The time taken for one whole cycle of an EAF heat, from the swinging of the roof for the first charge of scrap to the completion of fettling immediately prior to the next heat, is referred to as "the tap-to-tap-time". 26. The shorter the tap-to-tap-time to produce each heat of 22 M/T of liquid steel, the more heats per day, and, hence, the greater production over the course of a year. 27. According to SFI's case, the company, in the No-Scheme-World, on average would have needed only 23 minutes (0.38 hours) "useful" power-off time per heat. breakdown for that 23 minutes is:
Tap-to-tap-times for the 2 EAF's according to SFI's case are, therefore, as follows:
28. That means EAF 1 could have done 10.48 heats per operating day (24/2.29), and EAF 3, 13.41 heats (24/1.79). Thus, the two EAF's jointly, would have done 23.89 heats per operating day. 29. Based on SFI's 275.97 operating days, and 22 M/T of liquid steel per heat, that would mean a combined total for the 2 EAF's of 145,046 M/T of liquid steel per year (275.97 x 23.89 x 22). 30. The government's case was that 30 minutes "useful" power-off time was required for each heat. That 30 minutes is made up of:-
Tap-to-tap-times on the government case work out as follows:-
31. On that basis, EAF 1 would have had 9.96 heats per day (24/2.41), and EAF 3, 12.57 heats per day (24/1.91). Jointly, the two EAF's would, therefore, have done 22.53 heats per day on the government's case, so that, over the span of their 237.05 operating days in a year, SFI would have made 117,502 M/T of liquid steel (22.53 x 22 x 237.05). 32. From Exh SF81, one sees that, on average, EAF 1, over the period 2nd to 4th January 1982, took 3.45 hours tap-to-tap-time. With EAF 1 working at that speed and EAF 3 at its faster pace, SFI point out that it would have made over 100 000 M/T of good product per year. That, however, is without making any allowance for factors such as planned maintenance, and a mere three days' operations on one EAF is too narrow a foundation for drawing much in the way of inferences, in our view. 33. Mr Willcox's average tap-to-tap-time of 2.41 hours for EAF l in the No-Scheme-World is less than SFI, in fact, took on any of the 13 heats covered by Exh SF81, which, in this instance, indicates he was tending towards an approach favourable to SFI. 34. While actual power-off time - both "useful" and "non-useful" - can be seen for those three days on the one furnace, it would not be reasonable to make extrapolations from that. The clinching point is that SFI's calculations in Exh SF81 are based on 335 operating days per annum being available, whereas, in SFI75, SFI implicitly concedes there would, at most, have been 275.97 such days. 35. Before going on to examine the calculations leading to SFI'S figure of 275.97 and the government's figure of 237.05 operating days, we will first of all scrutinize the difference of 7 minutes between SFI's "useful" power-off time in the No-Scheme-World which amounts to 23 mimutes, and the government position that the correct figure should be 30 minutes. 36. That 7 minutes' difference is made up of 5 minutes for a third charge of scrap into the EAF, and 2 minutes more for the routine procedures of fettling, electrode adjustment and the like which occur on most, if not, every heat. 37. The evidence is plain that SFI's normal practice before the threat of resumption in November 1981 was to do 3 charges per heat, each charge taking SFI 5 minutes on average. That each charge on average took SFI 5 minutes was never made a live issue in the case, and, as the case proceeded on that assumption, we find as a fact that a charge in the No-Scheme-World would, on'average, have taken 5 minutes. 38. Arriving at a fair assessment of SFI's performance in either the Scheme-World or the No-Scheme-World has been rendered more difficult than it need have been on account of the unfortunate circumstance that SFI threw away nearly all its operating records on the occasion of physically vacating the dunk Bay premises in January 1987. 39. There does, however, reamin a memorandum dated 15th October 1979 (Exh. R12, p.57) from Mr Leung Po On, SFI's works' manager, which reveals that the practice at that time was to charge the EAF's three times. He also indicated that the total time taken for those 3 charges should not exceed 25 minutes. 40. There is also a memorandum, Exh SF123, dated 20th August 1980 from Mr Roy Leung, indicating that, if need be, there should even be a fourth charge. 41. As already indicated, three days' operating records for the period from 2nd to 4th January 1982 in relation to EAF 1 miraculously survive. There is an agreed translation and form of presentation of them in Exh SF81. They show that, of 13 heats, 2 required only 2 charges, while the other 11 heats needed 3 charges. 42. According to Mr Roy Leung, his men not trying very hard at that time, since they were not were not under any pressure then to produce more, as sales did not necessitate a faster pace, but he thinks they could have cut down to 2 charges per heat, if need be, to fulfil delivery orders. 43. Two other points were made by Mr Roy Leung to support his contention that his furnace crews, regardless of what had happened previously, could have managed with 2 charges by the time the threat of resumption occurred. 44. One reason was that, in 1981, charging baskets of what is known as the "split-shell" design, with a capacity of 16 cubic metres, replaced the previous baskets which held only 12 to 14 cubic metres. The second reason was SFI's acquisition of a "Best"-brand shear and an "alligator" shear for the scrap yard in 1980 and 1981, respectively, which meant that large pieces of scrap could thereafter be cut to manageable sizes, thus increasing the density of the charged scrap. The denser the scrap, the less volume it occupies. 45. In his report 30/01, dated February 1988, Mr Medley, SFI's steel expert-witness, made the general observation at page 10 in relation to charging that, "Depending on the nature of the scrap, that is weight to volume, a third basket may be required". 46. However, when he testified before the Tribunal, he was adamant that SFI's workmen would, normally, in the No-Scheme-World, have needed 2 charges only, to do a heat producing 22 tons of liquid steel. 47. From his evidence, it was unclear whether he knew that SFI's practice, at the time resumption was threatened in late 1981, was 3 charges. There was no suggestion that he had seen the material making up Exh SFI 81, the operational records for 2nd - 4th January 1982, before giving his evidence. 48. As all other records of that nature had been destroyed, and as he had never seen SFI's factory in operation, he was forced to rely on what Mr Roy Leung and other senior staff from SFI told him, plus his own knowledge and experience of what happens generally in mini-mills around the world. 49. As he put it, his object was to find out what was achievable or possible in the No-Scheme-World, rather than what had been achieved in the Scheme-World. 50. He described, in this context of what was achievable, how he would make a calculation, and ask Mr Roy Leung and Mr P.O. Leung (the works' manager) if they thought the result of that calculation was fair. If they did, he would treat that as being what was achievable. 51. An approach such as that does not strike us as a satisfactory way of seeking out what is truthful and reasonable. It is objectionable for the same reason that leading questions are objectionable. They suggest an answer to the person being questioned. 52. Using the approach he did, and with very little hard data on what SFI had actually done in the past, Mr Medley helped pave the way for a capacity claim which we regard as excessive, if, as he claimed, his opinion was formed on the basis of the 1982 installation. 53. When, in discussion with Mr Medley, Mr Roy Leung told Mr Medley he would arrange for 2 charges, Mr Roy Leung, in our view, was displaying a degree of optimism in no way justified by SFI's history. Moreover, we did not regard Mr Roy Leung as a wholly reliable witness. We will have more to say on his credibility in the context of how he came to claim that SFI, with its 1982 installation, in the No-Scheme-World, would have had the capacity to make 110,000 M/T of good product, despite he himself, (as well as others connected with SFI), having.consistently gone on record, at least until 1986/1987, to the effect that SFI's capacity was 100,000 M/T on the basis of SFI's installations in 1982, that being the last year SFI carried out any major improvement to its plant. 54. Because of our doubts concerning Mr Roy Leung's credibility, we are not inclined to attach much weight to his contention that the EAF crews, because they were not under any pressure, were deliberately taking their time during the period 2nd to 4th January 1982, to which Exhibit SFI 81 relates. We see no reason to doubt that Exhibit SF81 typifies the performance of SFI's furnace crews, and we find nothing in the evidence indicative of a future improvement from three to two charges for the EAF's. 55. Whether SFI would have needed two or three charges in the No-Scheme-World falls within the province of the steel experts - Mr Medley for SFI, and Mr Willcox for the government. 56. Both of them impressed us as highly knowledgeable about mini-mills. We felt there was generally nothing to choose between them when it came to matters of professional expertise. True, from their curricula vitaes, Mr Medley had looked at, and dealt with, more mini-mills than Mr Willcox, but it would be simplistic to suppose that an expert's knowledge about mini-mills is in proportion to the number dealt with. From the degree of familiarity Mr Willcox had with mini-mills, from his general knowledge of the steel industry, and because his speciality is process-engineering, we ultimately came to the view that, for the purposes of the present case,, he was as knowledgeable as Mr Medley on the mini-mill issues which arose. In addition to his own knowledge, Mr Willcox could draw on the expertise of his colleagues in the steel consultancy firm of W.S. Atkin. Mr Medley, too, could call on the pool of expertise in his firm, McLellans. 57. Because we regarded Mr Willcox and Mr Medley as having the same level of expertise for the purposes of the present case, we found ourselves constrained by the onus of proof to adopt Mr Willcox's, rather than Mr Medley's opinion, when there was a straight conflict of views between them without anything further in the evidence, or by way of inherent probability, to suggest that Mr Medley's opinion should prevail. 58. On behalf of SFI, it was urged upon us that, generally, Mr Medley's opinions should be preferred, since he had the edge over Mr Willcox, in that' Mr Medley had direct access to the Leung family, whereas Mr Willcox did not. We regard it as of no significance that Mr Willcox under cross-examination agreed that it.was an advantage for a steel expert to have direct access to the Leung family. Presumably, in making that answer, Mr Willcox took it for granted that the Leung family would have told the steel expert the truth, without any exaggerations in their answers. 59. As Mr Willcox was put forward only as an expert on steel, and not as a supposed expert on whether witnesses have told the truth, we do not consider that his agreement with SFI's counsel about the advantages of access to the Leung family is of any consequence. 60. It is for this Tribunal to decide questions bearing on credibility, and, on the view we take, Mr Willcox's agreement on the advantages of interviewing the Leung family is neither here nor there. 61. Had the Leung family been disinterested observers, Mr Medley's access to them might have added some weight to his opinions. As they had obvious purposes of their own to serve, we do not regard Mr Medley's access to them as an enhancing factor for his opinions. 62. In fact, Mr Medley's credibility sustained some damage as the result of Mr Roy Leung's access to Mr Medley. Mr Medley changed his opinion on the life of a concast from 15 to 20 years under pressure from Mr Roy Leung. 63. We also learnt from Mr Medley that Mr Roy Leung, not content with the 110,000 M/T's of good product, to which Mr Medley's calculations and opinions gave rise, wanted Mr Medley to go for an even higher capacity than 110,000 M/T under the 1982 installation. That time, Mr Medley did not succumb. 64. As we are of the view that the evidence shows Mr Roy Leung had consistently believed that SFI's capacity on the basis of the 1982 installation was only 100,000 M/T before Mr Medley came along, one can see the sort of genie Mr Medley let out of the bottle when he came up with his higher figure. 65. A further cause of diminished credibility for Mr Medley was his effort (in his report 30/04 page 62) to support his estimate of 23 minutes power-off time for SFI by reference to the power-off times for the mini-mills at Shearness Steel and Toshin Steel Co Ltd. Shearness, on average, needed 13 to 18 minutes power-off time, and Toshin only 10 minutes. As pointed out by Mr Medley himself, both of those mini-steel works had water-cooled panels, so needed less power-off time as there were no refractory bricks in their side-walls to undergo minor repair at the end of each heat. That said, we do not think it was helpful to the Tribunal to have the example of those 2 particular mini-mills quoted to us in the context of SFI's power-off time for the No-Scheme-World. 66. Shearness Steel, so we gathered, has the reputation of being at least amongst the finest, and is, perhaps, even the finest mini-mill in the world. Toshin Steel, too, obviously ranks very highly : it is used as a "shop-window" by Nikko Steel Ltd, a Japanese manufacturer of steel plant and machinery. Mr Roy Leung and his accompanying staff were obviously dazzled by Toshin Steel's performance when they visited Japan on a study tour in 1979. 67. SFI was clearly nowhere near the same class as either of those two mini-mills, and there was nothing to suggest it was ever likely to be. 68. Nor was it helpful to the Tribunal to have Mr Medley advance his racial theories on the steel-making characteristics of Chinese relative to Nigerians (See Mr Medley's Transcript, page 526 & 527). Raw data on the production capacities and achievements of various mini-mills in Taiwan (See Mr Medley's report 30/04 at page 96), and on Feng Hsin Iron and Steel Co Ltd in particular (See Exh SF 82), did nothing to increase our understanding of SFI's likely performance in the No-Scheme-World, since we know nothing on how the calibre of the management and workforce of the Taiwan plants compares with SFI's, nor do we know whether any of the Taiwan plants suffered the debilitating effects of a chronic cash-flow problem like SFI's. 69. While on the topic of Mr Medley's credibility, we noted that, somewhat obligingly, in his First Report 30/01 he described SFI's first concast machine as ``experimental". From that description one might be excused for imagining that SFI had a proto-type on a test-bed, but the reality happened to be that their first concast was an ordinary production model. 70. Only in one sense could that first concast be described as "experimental", namely, that SFI, with its trial and error approach, would have to experiment with the machine to see if they could get it to work. On that basis any new type of machine acquired by SFI was experimental. 71. Whether SFI would have managed with only 2, rather than 3 baskets of scrap for charging its EAF's in the No-Scheme-World, is a matter of expert opinion of the type where, on a conflict of views between the experts, Mr Willcox's opinion should prevail, as it is at least as inherently likely as Mr Medley's. 72. In fact, on this issue, there is positive evidence on why Mr Willcox's opinion should prevail. Not only was there a long history of SFI normally charging its EAF's three times per heat, but there was also evidence that SFI had difficulty getting scrap of sufficiently highquality to permit of 2 charges only. 73. There were 2 reasons, so we gathered, why SFI could not get scrap of the density it wanted. Firstly, and probably most importantly, SFI's financial state was so parlous at the time resumption was threatened that it could only buy from dealers willing to grant credit. That topic is ventilated in our' Section II entitled "Scrap Cost". 74. SFI had no choice other than to take whatever type of scrap the few credit-granting dealers happened to have. 75. On our calculations, SFI would not have had sufficient cash-flow to switch to cash payment for at least part of its scrap until Financial Year 1986/87, and for all of it in 1987/88 and beyond. 76. Now, for our second point. Once SFI was in a position to pay cash, it would, no doubt, have been able to shop around for the best scrap available on the market. Even then, it is far from sure that it would have been able to get all the high density scrap it would need to charge its furnaces regularly with two, rather than three charges. As Mr Len Leung said, you could only buy what was available. 77. For the foregoing reasons, we find, as a fact, that, in the No-Scheme-World, based on SFI's installation in 1982, fifteen minutes of "useful" power-off time would have been taken up by SFI making three charges, each on average of 5 minutes. 78. Thus, five minutes of the 7 minutes difference between SFI and the government over "useful" power-off time have been accounted for. 79. The remaining two minutes arise from Mr Medley's allowing 8 minutes on average for fettling, electrode adjustment and the like, whilst Mr Willcox said it needed 10 minutes. Judging from the column "Time End of Tap to 1st charge. Mins" on page 5 of Exh SF81, it is clear that, in early 1982, SFI was taking well over 10 minutes for this. Disregarding heat No. 17783/007, which took an abnormally long 72 minutes for fettling etc., (probably because it was necessary to change an economizer), the average time was 12.33 minutes. 80. Mr Medley's estimate of 8 minutes strikes us, even without recourse to the onus of proof, as, if anything, less likely than Mr.Willcox's 10 minutes, in the light of SFI's history, as revealed by Exh SF81. 81. Thus, overall, we find in favour of Mr Willcox's opinion that "useful" power-off time for SFI in the No-Scheme-World should be 30 minutes. 82. Attention can now be directed to how SFI and the government reached their respective totals of 275.97 and 237.05 operating days per year. 83. We now set out each side's position (subject to some minor modifications for ease of understanding) in tabular form:
(To complete the picture, we now extend the table to show each side's calculation for liquid steel, as referred to earlier in this Appendix, together with agreed yields)
84. Earlier in this present section, we made a passing reference to "availability/utilization" factors. 85. Understanding this case has been made a little harder than it need.have been for laymen such as ourselves by Mr Medley labelling his factors in terms of "availability", and Mr Willcox opting for "utilization". They could just as easily have both used the same word, without either of them weakening his argument. 86. "Availability" when used in conjunction with "factor" had the meaning assigned to it by Mr Medley of a period of time when the works was available for production, calculated without any separate allowance for planned maintenance at any stage. Thus, for example, Mr Medley's line 6 in the table above shows the figure of 298.35 days, calculated by applying Mr Medley's availability factor of 85% without any separate allowance for planned maintenance anywhere in Mr Medley's column. Mr Medley told us that planned maintenance, which both he and Mr Willcox agree takes 16 days, is already included in Mr Medley's 85% availability factor. 87. By contrast, when Mr Willcox adopted the word "utilization", it implied that the 16 days' planned maintenance was not already embraced by, say, the 85% utilization factor of his line 5, but amounted to a separate deduction, as can be seen in line 3 of Mr Willcox's column. From the point of view of language generally, we regard this as a somewhat arbitrary use of the word "utilization", but it is one by which we have been required to abide for the purposes of the present case. 88. A reconciliation can be brought about between Mr Willcox's 85% utilization factor and Mr Medley's 85% availability factor by adding 4% (representing the 16 days planned maintenance) to Mr Medley's 85% availability factor. Thus, the 298.35 operating days of Mr Medley's line 6, can just as easily be expressed in terms of 89% of the 335 days in Mr Willcox's line 4, as by treating them as 85% of the 351 days of Mr Medley's line 4. 89. The differences between the parties, as shown by our above table, can quickly be narrowed down. 90. Mr Willcox's contention, reflected in line 3, that 16 days planned maintenance should be allowed for, prior to applying the 85% utilization factor, is every bit as plausible to us as Mr Medley's contrary assertion, so, the onus of proof resting with SFI as it does, we find in the government's favour in respect of those 16 days. 91. By the same token, there was nothing to persuade us that Mr Medley's opinion the continuous casting machine does not deserve its own utilization factor in line 9 is any more likely to be right than Mr Willcox's view that a 90% factor is justified. Mr Willcox's explanation that the continuous casting machine has its own problems with mechanical and electrical breakdowns, breakouts, and other unplanned stoppages over and above those provided for in line 5's 85% utilization factor, did not strike us as any more unreasonable than Mr Medley's denial. 92. The real tussle boiled down to what was comprised in each party's 92.5% scheduling factor at lines 6 and 7. 93. As SFI's written submission, C.C.S. Section 2, para. 3.1 at page 12 so rightly noted, there was a degree of overlap in what, in effect, were the factors for availability/ utilization and scheduling, as well as for the power-off time. Neither would we dissent from SFI's view that it was "... very difficult to identify the extent to which there was really agreement on the proper allowance for each of these factors". 94. Much-of the argument had a metaphysical quality to it, and nowhere more so than in SFI's contention that its 92.5% scheduling factor was not the same as the government's 92.5% scheduling factor. 95. SFI'S 92.5% scheduling factor had its genesis in Mr Medley's first report 30/01 at page 31 under the heading. "Continuous Casting". (In the interests of simplicity, we accede to SFI's request that we should disregard its earlier reference to the topic of scheduling at page 23 of that report under the heading "Steel Melting"). 96. At page 31 of his report (30/01), Mr Medley explained that, on the basis of his availability factor of 85%, there was sufficient time to cast 22 tons of liquid steel in the concaster on 147 occasions in the course of a week of 143 hours (168 hours x 85%), whereas the combined output of the two EAF's over the same period amounted to only 142.32 heats of that size. (At page 31 of 30/01, Mr Medley calculated the number of heats as 140, but that requires revision, upwards, in the light of our finding of quicker melt-down times. The arithmetic now is:
97. It was conceded on behalf of the government there would have been sufficient time for SFI to put those 142.32 heats, each of 22 M/T of liquid steel, through the concaster by single casts as opposed to sequence casts. Assuming a casting speed of 3 metres per minute as Mr Medley did in 31/01, 22 M/T of liquid steel can be cast in about 33 minutes which, together with 25 minutes restranding time for the dummy - bar, means a single cast should be completed in 58 minutes. 98. Still at page 31 of Mr Medley's report 30/01, he went on to observe, "The number of casts per week could be increased by sequence casting, but there is however the difficulty of scheduling the two arc furnaces operating on different tap-to-tap-times to ensure matching with the one continuous caster. Experience has'shown that unscheduled delays in melting, handling and casting can reduce output by between 5% - 10% over and above the availability factors, hence the arc furnace liquid steel output should be reduced by say 7 1/2% ..." 99. In his report (Document No. 45, first version, at page 21), Mr Willcox said that the scheduling problem between the two mismatched EAF's - EAF 1's transformer was 8,250 KVA while EAF 3's was 12,000 KVA - was such that casting was limited to the rate of the slower of the two. Leaving aside, momentarily, Mr Willcox's point about the casting rate being limited to the slower of the two mismatched furnaces, the scheduling factor described by Mr Willcox in the original page 21 of his report, Document No. 45, does not sound too different from Mr Medley's original description in 30/01 at page 31. 100. On the point about casting being limited to the rate of the slower of the two EAF's, Mr Medley in his report 30/04, page 4 et seq., clearly demonstrated that Mr Willcox was wrong on that, and, since then, Mr Willcox has modified his view by saying (at page 000022 of Document No. 45) that it would only be initially that the EAF's would be limited to producing at the rate of the slower of them, by alternate castings. That would result in a scheduling factor of 88.3% for 1981/2, and then, with gradual improvements, would be increased to 92.5% by 1986/7, the year when, according to Mr Willcox, SFI would have completed its learning curve and have reached its actual production capacity o?100,000 M/T. 101. Two points can be cleared out of the way immediately. One is that we disagree with Mr Willcox's five year learning-curve. The four year learning-curve he proposed in his earlier report, when his opinion on full capacity was 93,000 M/T, strikes us as appropriate, too, for Mr Willcox's modified capacity of 100,000 M/Ts. Of this, more will be said, later in this Section, when we come to deal with the related topic of production build-up. The other point is that we consider Mr Willcox's micro-adjustments of his scheduling factor amount to an unrealistic attempt to be precise in an area where a broad brush approach strikes us as more appropriate. 102. If there is to be a scheduling factor, as the government contends, then, in our view, it should be a straight-forward 92 1/2%. 103. In relation to scheduling the EAF's and the concaster, the principal point at issue between the parties is whether, as SF1 contend, scheduling is, in effect, irrelevant, since the continuous casting machine has sufficient capacity to take all the liquid steel from the EAF's by single casts, or whether, as the government maintains, two mismatched furnaces feeding the one continuous casting machine still give rise to a scheduling problem, notwithstanding the fact that the number of single casts available in any period exceeds the number of heats. 104. According to SFI, scheduling only arises as a problem if there is sequence casting, which is to say, as soon as one ladle of steel has finished pouring into the concaster, a ladle of steel from the other EAF immediately starts pouring, without any break, thereby avoiding the loss of the twenty-five minutes of the concaster's time required for re-stranding the dummy-bar. 105. When Mr Medley next spoke about his 92.5% scheduling factor, it was in different terms from the passage already quoted from page 31 of Document No. 30/01 under the heading "Continuous Casting". 106. From page 175, starting at line 15, of his transcript, there is the following from Mr Medley's re-examination:-
107. That is a far cry from what he said in his Document No. 30/01 at page 31, since he now, in effect, seems to be saying his availability factor and his scheduling factor cover the same ground, namely, unplanned stoppages. 108. We now quote what he said in his report 30/01 at page 31:-
109. In a transmutation which the Schoolmen might have admired for its opacity, Mr Medley came up with a third version of his scheduling factor.. This is to be found in Exhibit SF175C, line 10, column H where the "scheduling factor of 90 - 95%" is said to be "for unscheduled non-coincident delays on caster and other steel plant and equipment". That third version gives no hint of "the difficulty of scheduling the two arc furnaces operating on different tap-to-tap-times to ensure matching with the one continuous caster", referred to in Mr Medley's first version. 110. Mr Medley has tried to distance himself from the notion that E.A.F.'s with different tap-to-tap-times give rise to a scheduling problem when they are coupled to a single continuous casting machine. 111. Is SFI right when it now seeks to say that the type of scheduling problem arising from mismatched EAF's and the one continuous casting machine can only occur in a sequence casting situation? 112. We do not think it is. As explained by Mr Willcox, the scheduling problem, stemming from mismatched EAF's and the one continuous casting machine, arises from what he described as the "bunching" of heats when both furnaces reach the stage of being ready to pour at more or less the same time. Single casting would not render SFI immune from that problem. With, on SFI's case, 142.32 heats per week, and with the concaster available for 147 single castings, it seems to us that Mr Willcox was right in saying that there was very little of what he described as "balancing accommodation" between heats. 113. Moreover, there is also some evidence from Mr Medley himself on how there can be a scheduling problem even with two precisely matched E.A.F.'s (See Mr Medley's transcript, pages 602 and 603). 114. The 92 1/2% scheduling factor was made necessary according to Mr Willcox by three possible events. 115. Firstly, there might be interferences between one of the two charging cranes and the other, or between one of the two casting cranes and the other. 116. Secondly, SFI's melting materials store was separated from the casting bay by the melting bay. That contrasted with the better practice of more modern mills which had overhead bins for such materials in the.casting bay. BFI's system increased the risk of the right melting material not being available in the right place in the right quantity at the right time 117. Thirdly, there was only one ladle pre-heating station. 118. We must say that, by themselves, we find those three possibilities enumerated by Mr Willcox a trifle on the thin side to justify a reduction in the number of available days by as much as 7 1/2%. 119. As already indicated, we have accepted Mr Willcox's testimony to the effect that an 85% utilization factor should be applied to SFI's melt-shop, and a separate 90% utilization factor for the concast. We also accepted what he told us about those factors being empirically based. 120. At this point, it is convenient to record our finding about the standard at which BFI's plant and machinery could reasonably have been expected to be maintained in the No-Scheme-World, since maintenance will obviously have a bearing on utilization/availability factors. 121. According to Mr Roy Leung and Mr Ho, the engineer in charge of maintenance, SFI's plant and machinery was particularly well maintained. On that basis, it was argued on SFI's behalf that it would have had a better availability/utilization rate than the 85% for the meltshop and 90% for the continuous casting machine, given by Mr Willcox on the basis of what happened in other mini-mills. 122. We do not doubt that SFI's plant and machinery was adequately maintained in the Scheme-World (except perhaps in the last few months running up to the cessation of manufacturing in August 1986), but there is nothing on the evidence to support the notion that it was any better than average for the industry. 123. As far as we can make out from the evidence, no records were kept of break-downs of plant and machinery, and even if they had been, it would not have been SFI's style to analyse them to show the amount of down-time, say, per week, or per month, caused by break-downs. 124. Mr Ho actually sounded rather offended when it was put to him in cross-examination he should have kept records on the down-time of the machines he maintained. 125. In relation to Mr Ho's credibility generally, we gained the impression that he was an honest, but somewhat confused witness. At one point in his evidence he claimed that he was even more important than the factory manager in relation to production yields, but then at another point went on to say he would not have been a suitable person to take charge of production. In our Section VI, PLANT AND MACHINERY we comment on the unsatisfactory nature of his evidence concerning spares. 126. Our overall impression of Mr Ho's evidence was that little weight could be attached to it. 127. In the light of what little we know about the maintenance of SFI's plant and machinery, we see no reason to deviate from Mr Willcox's utilization rates of 85% and 90% for the meltshop and continuous casting machine, respectively. 128. A utilization factor of 85% for the meltshop and 90% for the continuous casting machine produces an overall utilization rate of 76.5%. Such an overall utilization rate was exemplified, according to Mr Willcox, by Maia, a minimill in Portugal of which he had personal experience. There is some controversy about Maia's actual utilization rates, but, suffice it to say, they were probably somewhere around 75% to 80% during the period 1984 to 1986. See Exhibit R48. We do not think it necessary to make 'a separate finding of fact about Maia's precise utilization rates, since it is a collateral issue. We think that Mr Medley was right and Mr Willcox wrong about electrode changing being part of useful power-off time. We accepted what Mr Willcox told us about Maia being built in the 1970's, and that it ranked above SFI in the hierarchy of mini-mills in terms of up-to-date technology. 129. It struck us as a valid point when Mr Willcox drew our attention to the circumstance that Maia and various other named mini-mills, including Hong Kong's Shiu Wing Steel Ltd, had the advantage of a single EAF, served by a single charging crane, feeing a single continuous casting machine, served by a single ladle crane. All support services such as electricity and water would be dedicated to just that one unit. 130. His 85% utilization rate for the meltshop and 90% for the concast were based on this sort of single unit mini-mill, he explained. 131. From there he went on to assert that, if a second unit were added, such as, for example, a second charging crane or a second EAF, that introduced a scheduling problem which would reduce the utilization rate. 132. In principle, Mr Willcox's argument impresses us as reasonable, but we have hesitated over whether the specific potential interferences he pointed to in SFI's case merited an additional scheduling factor of the magnitude of 92.5%. 133. The strongest argument in favour of allowing Mr Willcox's full 92.5% scheduling factor along with his utilization factors is that it is more consistent with a bottom line under the 1982 installation of 100,000 M/T's of rebar, a figure for which there is powerful corroboration in the evidence of SFI's directors and advisers before Mr Medley, and Deloitte Haskins came upon the scene in late . 1987. 134. Consistently, from as early as 1972 when 51% of SFI's shares were bought by New World, Mr L.Y. Leung, the Managing Director of SFI, has been saying that his expansion programme, involving an additional EAF, (which became EAF 3), a continuous casting machine, an additional re-heat furnace (which became R.H.F."C"), and an additional rolling mill (which became R.M.3), would result in a doubling of capacity to 100,000 M/T's of rebar. 135. Mr Roy Leung, who was put in charge of the Expansion Programme when he became a director in 1975, consistently followed the same refrain as his father. 136. On the occasion in November 1981 when Messrs. L.Y. and Roy Leung.met overnment officials to discuss the implications of the government's threat of resumption at Junk Bay, Mr L.Y. Leung stated that the then present capacity of his mini-mill was about 100,000 tons per annum (Exh SF 94 page 3). 137. A letter dated 10th February 1983, from the solicitors employed by SFI for the conduct of its case, to the government, stated on page 3 (Exh SF 94 page 29) that SFI's "production capacity has increased 33 times in the last 32 years from an output of 3,000 tons per annum in 1951 to the present output of 100,000 tons per annum.. At page 5 of the same letter, SFI's solicitors explain how their client, in December 1981, had got an estimate from Mannesman Demag A.G. of Germany for a new steel plant capable of producing 100,000 tons of steel per annum. 138. That letter has also to be seen in the context of SFI having retained Jones Lang Wooton ("J.L.W."), the internationally-known estate agents and valuers, in early 1982, to work on the financial implication arising from the threat of resumption and to assist with the preparation of SFI's claim for compensation against the government. It would not be unreasonable to presume that there was some coordination between SFI's solicitors and JLW, either directly or'through the directors of SFI. 139. It would be difficult to imagine more powerful corroboration for SFI's capacity being 100,000-M/T's under the 1982 installation than the particularised claim lodged with the government under the Ordinance by J.L.W., on SFI's behalf, by a letter dated 30th October 1986 : Exh SF 95 page 51 et seq. 140. JLW had the following to say in paragraph 4 of their letter:
141. The calculations for past profit estimates and future profit forecasts referred to in paragraph 4 of JLW's letter can be found in Exh SFI 98, which comprises a letter dated 30th October 1986, to the directors of SFI from Price Waterhouse who were SFI's auditors (as well as New World's) at all material times up to 1987, when Price Waterhouse took the initiative in discontinuing as SFI's auditors on the basis of some imagined conflict of interest and duty in the event of any of its staff being called as witnesses in the present reference. We accept that it was nothing of SFI's doing which led to Price Waterhouse bowing out from the case. 142. That letter of 30th October 1986 from Price Waterhouse to SFI's directors had the following content :
143. Taking up the theme of " .... the accuracy of the notes and assumptions made by the directors", one turns to the fifth page of Exh SFI 98 for the "Explanatory Notes And Assumptions" accompanying "the estimates and forecasts". 144. Under "Explanatory Notes And Assumptions" appears the following "The past profit estimates and future profit forecasts have been prepared on the following assumptions:-
145. With Price Waterhouse out of the way by virtue of Price Waterhouse's own initiative, but with J.L.W.dismissed by SFI, SFI was free to seek out new accountants and valuers, unburdened by exhaustive knowledge of SFI's past. 146. It would have been helpful to the Tribunal to hear J.L.W.'s explanation of why SFI's directors had consistently described SFI's capacity at the time of the threat of resumption as 100,000 M/T's, but now say it should be 110,000 M/T's under the 1982 installation which basically remained unchanged until SFI actually vacated Junk Bay in January 1987. 147. SFI's capacity claim of 110,000 M/T'S per annum, based on the 1982 installation, has an aura of recent invention about it. 148. We do not believe for one moment Mr Roy Leung's contention that he knew all along that SFI's capacity was more than 100,000 M/T's, but he did not like to contradict what his father said at the meeting with government officials on 19th November 1981 (Exh SFI 94, p.3 to 5), and how, thereafter, Mr Roy Leung had to stick to 100,000 M/T's for the sake of consistency with what Mr L.Y. Leung had said on 19th November 1981. 149. A clear instance of Mr Roy Leung not telling the truth occurred, in our opinion, when Mr Roy Leung sought to explain how it came about that his father, the managing director, who was very much in control of the company so we were told, was under the mistaken impression SFI's capacity was only 100,000 M/T's in November 1981 when the allegedly true capacity, known to Mr Roy Leung then, was in excess of that. 150. It was obvious to us that Mr Roy Leung was making up a story as he went along when he testified that his father had not been informed of improvements for SFI's mini-mill stemming from what Mr Roy Leung and other staff had learnt in 1979 on a study tour of Toshin Steel Co Ltd's mini-mill in Japan. Credulity was stretched way beyond breaking point by that answer on two grounds : firstly, it was inherently improbable that, during a period stretching from 1979 to November 1981, the managing director, Mr L.Y. Leung, would not have been given the good news by his son and fellow-director, Mr Roy Leung, that SFI's capacity was going to be increased beyond 100,000 M/T's on the basis of what was learnt on the Toshin trip,and, secondly, the so-called "Toshin improvements", such as, for example, better scrap management, either had not been implemented, or, even if they had, as with, e.g. the split-shell charging baskets, had not resulted in any significant increase in production by the date of the meeting with government in November 1981. 151. A line of argument put forward on SFI's behalf has been to the effect that, although SFI's target at the outset of its expansion programme dating from the early 1970's was a capacity of 100,000 M/T's, there have, however, been technological developments in mini-mills since then which have opened the way for SFI to think in terms of a larger capacity for its 1982 installation. 152. That argument cuts little ice, since the evidence suggests that at the time when SFI made its estimate of 100,000 M/T's at the beginning of its Second Phase' of Expansion at Junk Bay, SFI was not in fact technically capable of making high-tensile rebars consistently, and had not grasped how difficult it would prove for them to do so, on account of the more advanced level of technology and skill required. 153. From the time in 1978 when SFI got its continuous casting machine working and installed a liquid oxygen system - on the evidence, the two vital steps enabling SFI consistently to make high-tensile rebars of the quality required by British Standard 4449 which applied in Hong Kong - until mid-1982, when the shadow of resumption's effects started to intensify, SFI got stuck in a production range of about 45,000 to 55,000 M/T's per annum (see Exh R31 and our "omnium,gatherum" table at the end of this section, entitled "The Tribunal's Table"). 154. We do not doubt that from 1978 onwards, SFI's plant at Junk Bay already had a capacity of 100,000 M/T's of high-tensile rebar, but SFI failed for the next four years or so to increase production significantly, the main reasons being, in our view, that SFI had not mastered the necessary technology, and it was in such a mess financially that its manufacturing costs (which included scrap bought on credit at inflated prices) exceeded its sales' prices for its rebars. 155. Such additional improvements as SFI made from 1979 onwards were, on the view we take of the evidence, peripheral, and did not add to SFI's capacity of 100,000 M/T's. 156. In our considered view, it would fly in the face of common sense to hold that, with the installations it had in 1982, SFI's actual production would have exceeded 100,000 M/T's. If anyone knew SFI's production capacity, it surely must have been its directors who actually carried out the Expansion Programme, and who had actually to make the installation produce, rather than Mr Medley who comes along to theorise in 1987 after the mini-mill had been dismantled. 157. Against a background of all the evidence in the case, we are prepared to accept that, interferences between cranes, readiness of melting materials for adding to casting ladles, and the circumstanceof only one ladle preheater rather than two, justified a scheduling factor as high as 7 1/2$. When we speak of interferences between cranes, what we have in mind is one charging crane blocking the other charging crane, or one casting crane blocking the other casting crane, such blockages arising from poor organization. 158. We have not overlooked Mr Medley's Exh SF195 which shows that, on average, each casting crane only ever had to perform one manoeuvre every 83.4 minutes. Averages can be misleading, a popular example being the case of someone drowning in a pond on average three inches deep : it was just the victim's misfortune to fall in a hole there which happened fallacy of water six feet deep. We think a fallacy of a similar nature can arise from thinking in terms of averages in relation to crane movements. We accepted what Mr Willcox told us about the cranes moving slowly and how in practice interferences do arise. 159. Likewise, we felt the vast exercise, (See Exh SFI 176A and 176B) mounted on SFI's behalf, to show how few, on average, were the likely traffic movements within the Junk Bay premises in the No-Scheme-World did nothing to refute Mr Willcox's point in relation to the availability of the melt-shop that, with the scrap yard choc-a-block with sufficient scrap to produce 100,000 M/T's of rebar per annum, there could be traffic congestion on a scale which would make it less likely that SFI would be getting the right scrap, in the right quantities, at the right time, to the right EAF. 160. Mr Medley and Mr Willcox both fully participated in a vast arithmetical joust together, the results of which, we feel, have an artificiality about them which should not be allowed to deflect us from a bottom line according with commonsense. 161. Another reason advanced by Mr Roy Leung as to why SFI's mismatched furnaces did not justify Mr Willcox's 92 1/2% scheduling factor for poor synchronisation was that SFI, in the No-Scheme-World, would always have been able to hold a ladle of steel, ready to cast, for up to one hour, aided by a lining of Rosaki heat-retaining bricks in the ladle and a sprinkling of rice husks on top. 162. Mr Willcox doubted that time on the score of heat-loss from SFI's relatively small ladles, and, on that basis, he thought a more likely maximum time for SFI to hold its ladles ready for casting was 40 minutes. 163. We found Mr Willcox's evidence about the time a ladle of liquid steel could be held as credible as Mr Roy Leung's. 164. Of course, the less time the liquid steel could be held in the ladle, the greater the'likelihood of scheduling problems between the EAF's and the continuous casting machine. 165. A further point put forward in SFI'S favour was that it had already had the parts for a third strand on the continuous casting machine which, once installed, would have reduced the time needed for concasting by one third, but, it was conceded on behalf of SFI, that would not benefit its operations until melt-down times on the EAF's were reduced, and, hence, production capacity increased by, e.g., water-cooled panels on the EAF's or the addition of oxy-fuel burners to the EAF's. 166. All this (or at least, all of it except, perhaps, for the Rosaki heat-retaining bricks for casting-ladle-lining, and the rice-husk technique for trapping heat in the casting ladle) was surely known by Mr L.Y. Leung when he voiced the view to government officials at the meeting of 19th November 1981 that SFI's then present production capacity was about 100,000 M/T's per annum. Certainly, from 1982 onwards, until Mr Medley's arrival on the scene, absolutely everything, including Rosaki bricks and the rice-husk technique, must have been known to Mr Roy Leung on every material occasion when, without contradiction from himself, he permitted the same information to be repeated to the government, namely, that SFI's capacity under the 1982 installation was 100,000 M/T's. 167. A somewhat minor point tending to show that 100,000 M/T's rather than 110,000 M/T's was SFI's likely capacity under the 1982 installation is to be found in what we regard as a concealed availability factor contained within Mr Medley's evidence. 168. Applying the availability factors explicitly put forward by Mr Medley, he arrives at an answer of 126,799.21 M/T's of good product per annum. As part of what Mr Medley describes as his conservative approach, he reduces that to 110,000 M/T's per annum. 169. That reduction, however, turns out to be an allowance for such possibilities as a major plant break-down or a strike. In substance that is, in effect, another availability factor. We fail to see how there is anything essentially conservative in reducing that figure of 126,000 to 110,000 M/T's when it represents specified contingencies. 170. On the basis of SFI's 1982 installation, we conclude from the foregoing that SFI's maximum annual capacity in the No-Scheme-World would have been 100,000 M/T's of rebars. 171. The model agreed to by the parties is that the Tribunal is to determine SFI's capacity in the No-Scheme-World on the basis of the 1982 installation. Despite that agreement between the parties, we do not think that, in assessing SFI's loss or damage, we are precluded from taking into account improvements SFI might have made from 1982 onwards. A whole list of possible improvements was cited to us on SFI's behalf, including water-cooled panels, oxy-fuel burners, Tempcore, an oxygen-making plant, and 50mm rolls, as well as the techniques of hot-charging and sequence-casting. 172. At the time the "shadow" fell in late 1981, SFI was totally dependent on New World for financing. It is far from clear that New World would have been willing to make any further capital outlays, even without the "shadow". Rebar prices remained at an abysmal level until 1983/4 when there was a two year interlude of improved prices until 1984/5, after which prices drifted down again until 1987/8. The "shadow" coincided with a general down-turn in Hong Kong's economy which lingered until about 1986. 173. On our view of the No-Scheme-World, SFI would have continued to make losses and require further loans from New World until 1986/7 (see D.H.S.'s Revised Appendix II, of 6th March 1992 - a copy of which is produced opposite), the last time it had made a profit being 1974/5. With the Second Phase of SFI's Expansion Programme at Junk Bay completed by 1981/2, the continuing absence of profitable operations would have made it reasonable for NWD to look askance at requests from SFI for any further substantial sums for capital out-lays. Despite Mr Stewart Leung's reassuring words in the witness-box, we have grave doubts whether New World would, in fact, have been willing to make available, say, a loan to'SFI of the $20 million required for Tempcore, the fact that it would have resulted in a beneficial Net Present Value notwithstanding. 174. We think it more likely that SFI would, generally, have had to wait until it generated sufficient positive cash-flow for itself in the No-Scheme-World before it could invest in any further improvements requiring sizeable sums of money. Water-cooled panels would, however, in our view, have been an exception to that. 175. Because of the relatively small amount of money involved - $1.5m - and the vastly disproportionate advantages to be reaped - a saving of about $2.4m per annum
on refractories, and substantially increased capacity per heat due to the increase of the internal diameter of the E.A.F.'s which would no longer need a lining of refractory bricks - we do not doubt that N.W.D. would have made the necessary $1.5m available for whenever SFI decided it wanted to go ahead with that improvement.' When precisely that would have been is difficult to pinpoint. various factors would have come into play on the timing of that improvement. For example, the water-cooled panels would more likely than not have been postponed until after the third strand was added to the continuous casting machine, since there is little point in adding capacity for liquid steel unless there is a corresponding increase in concasting capacity. Another factor would be how far the work-force had progressed along the learning-curve. There would be no point in rushing to get water-cooled panels fitted until the workforce had developed the ability to cope with a higher volume of liquid steel. 176. The actual installation of the water-cooled panels would have interrupted melting for no more than one or two weeks. 177. As Mr Roy Leung was somewhat vague on his time-table for further improvements - at page 883, line 16 et seq of his transcript, for example, he referred to improvements in the No-Scheme-World being spread over two or three years - we cannot avoid an element of arbitrariness in allocating a date. The date should avoid being so soon that it might confer an unfair advantage on SFI, but, on the other hand, it should not be so far ahead in the future that SFI suffers an unjust deprival. 178. Neither side will be dealt with unfairly, in our view, if we nominate the commencement of Financial Year 1985/6 as the time by when SFI would have modified its EAF's by the addition of water-cooled panels. By then, SFI would have completed its learning curve. 179. With water-cooled panels, we are satisfied that, commencing with Financial Year 1985/6, SFI's capacity would have risen to a level where it could, without too much difficulty, make 110,000 M/T's of good product per annum. 180. Whether SFI's EAF's with water-cooled panels would have had the full 30 M/T's of liquid steel capacity Mr Medley claimed for them and whether SFI would have been able to produce at that level is something we feel it is impossible to decide on the evidence before us. We wish to avoid any pretence of spurious accuracy. Nor need we decide how much faster a heat could be done once the EAF's had water-cooled panels, one of the advantages of which is to allow longer arcs, thus reducing melt-down time. Availability/utilization would also improve as there would no longer be any refractory bricks in the EAF's requiring minor repairs between heats. 181. Instead of having to re-brick the furnace every three or four hundred heats, water-cooled panels can continue without any comparable attention for about 5,000 heats. 182. A further advantage of water-cooled panels is that less electricity per metric ton of scrap is needed to achieve melt-down. 183. Suffice it to say that, commencing 1985/6 in the No-Scheme-World, SFI would have had no great difficulty producing at the rate of 110,000 M/T's per annum. Further, we do not doubt that SFI's capacity margin above 110,000 M/T's would be such that SFI would have been able, in 1985/6, to rise directly to 110,000 M/T's from the figure of 100,000 M/T's capacity we found for 1984/5, without any need to prolong the learning curve. 184. From the above, it is clear that SFI would have garnered enormous rewards from water-cooled panels for a relatively tiny investment. By 1985/6, the year by when we believe SFI's EAF's would have been operating with water-cooled panels, we are satisfied that, in Mr Roy Leung's words, "everything would have come together". 185. The increased volume of liquid steel would have been accommodated by the continuous casting machine which would have had a third strand. 186. Sequence-casting would have been the rule, rather than the exception, by then. 187. SFI would, by then, have been deriving the full benefit from raising the overhead crane rail of the casting bay in 1980, and cutting away part of the continuous casting machine's platform for ladle manoeuvres in 1981, as preludes to sequence-casting. 188. By claiming that SFI's plant and machinery had prematurely attained a capacity of 110,000 M/T's in 1981/2, and that SFI would actually produce at that level from 1983/4 onwards, SFI has confronted the Tribunal with a problem. 189. The case was fought on the basis of capacity under the 1982 installation. For the reasons we have given, we do not think that SFI could have made more than 100,000 M/T's of good product under the 1982 installation. 190. However, for a modest out-lay of $1.5 million on water-cooled panels, SFI would, on the view we have taken, have actually been able to make at least 110,000 M/T's of good product from 1985/6 onwards. 191. Despite the case having been fought on the basis of SFI's 1982 installation, we do not think we have to ignore the likelihood of future improvements to that plant. 192. Three courses are open to us. 193. One would be to base SFI's cash-flow on production of 100,000 M/T's, but adopt a discount rate low enough to take into account the probability that SFI's production would have increased to 110,000 M/T's for 1985/6 and beyond. 194. A second course would be to grasp the nettle, and acknowledge that, despite the agreement between the parties about basing production on the 1982 installation, SFI's cash-flow from 1985/6 onwards would have been commensurate with production of 110,000 M/T's for that year and onwards. 195. That is the course we prefer, and, accordingly, adopt a cash-flow based on sales of 110,000 M/T's, rather than manipulating the discount rate with a "fudge factor" of the type Brealey and Myers counsel against in Principles of Corporate Finance, (3rd Ed.), page 188 et seq. 196. A third course, perhaps arguable on the evidence, would be to disregard the post-1982 improvement of water-cooled panels we now believe SFI would have made, and limit ourselves to capacity under the 1962 installation. The line of thinking would be that SFI, in the present adversarial proceedings, gambled and lost : it took a chance before us by staking everything on an unduly high capacity claim under the 1982 installation in the hope of persuading the Tribunal it could race up to 110,000 M/T's of production by as early as 1983/4. To adopt that approach would not bring about a just result, in our view. Just because SFI exaggerated its claim, and a few lies were told in the witness-box, should not mean that SFI is denied full compensation for such loss or damage as we are satisfied it sustained. 197. Attention can now be turned to SFI's build-up to capacity. We have already made mention of how SFI's actual production hovered within the relatively narrow band of between, approximately, 45,000 to 55,000 M/T's per annum over the years spanned by 1978 to 1982. 198. Until about mid-1981, SFI was drifting along in the doldrums with yields, scrap to ingot/billet, around the 83% mark, and from rolling, around, say, the 86% mark. (See "The Tribunal's Table" at the end of the present Section). 199. According to Mr Widdicombe, there is an explanation for those relatively modest yields over the period 1978 to 1980, namely, that SFI, during that time, was trying to master the technique of making high-tensile rebar, which is considerably more difficult to make than mild steel rods. Mr Widdicombe was only being very approximate with his dates, and we think the figures show that, roughly speaking, mid-1981 was the time SFI really got the knack of making high-tensile rebar. 200. Starting from August 1981, the yield from both scrap to billet, and billet to bar, soared. 201. For the purposes of the present case, SFI's agreed yield, scrap to billet, in the No-Scheme-World is 86.5% (92% scrap to liquid steel x 94% liquid steel to billet). SFI, in the Scheme-World, did at least as well as that (and, frequently, considerably better), virtually every month from August 1981, onwards. 202. While scoring around the 88% mark for rolling billets to good product from August 1981 onwards, SFI, in the Scheme World, did not start hitting above the 90% mark until February 1982, but, after that, came close to, or exceeded the agreed percentages for the No-Scheme-World of 92% (1982/3), and 93% (1983/4 and onwards). 203. That period from 1978 to mid-1981 when SFI was learning how to make high-tensile rebar coincided with the time SFI's staff were applying themselves to learning how to operate the continuous casting machine which was commissioned in October 1978. 204. A hypothesis has been put forward, on SFI's behalf, that, during the period 1978 to mid-1981, SFI had been readying itself for "take-off", and, with the dramatically improved yields from mid-1981 onwards, would have swiftly climbed to actual production of 110,000 M/T's of rebar. 205. Even on the SFI scenario, the take-off for the No-Scheme-World is not quite vertical, but proceeds, by way of a modest increase from rolling, of 3,393 M/T's to 55,934 M/T's for 1981/2 (when actual production from melting was 50,012 M/T's and from rolling, 52,541 M/T's), next, 86,000 M/T's for 1982/3, and culminating in 110,000 M/T's for 1983/4 and thereafter. 206. According to Mr Medley, SFI had finished its learning-curve on how to make high-tensile rebars by mid-1981, the time when, approximately, we have pin-pointed the attainment of dramatically improved yields. 207. As the learning-curve had been completed, the reason, so Mr Medley told us, for SFI not doubling its production overnight was that SFI's staff needed time to adjust to "the increased pace of working". Those words from Mr Medley in inverted commas, so we feel, are a euphemism for further progress along the learning-curve. 208. Mr Willcox posits a learning-curve for the No-Scheme-World where, after "take-off", altitude is gained more gradually. There is no change in Mr Willcox's model from the Scheme-World to the No-Scheme-World in 1981/2, but, for 1982/3, production for the No-Scheme-World becomes 75,000 M/T's (compared with 37,759 M/T's in the Scheme world), 85,000 M/T's in 1983/4, 93,000 in 1984/5, 98,000 in 1985/6 and 100,000 in 1986/7. 209. In his First Report, where Mr willcox found the maximum capacity of SFI's plant and machinery to be 93,000 M/T's per annum, he portrayed SFI as building up to it by steps of 75,000 M/T's in 1982/3, 85,000 M/T's in 1983/4 and then, finally, the 93,000 M/T's maximum in 1984/5. 210. We sympathise with SFI's complaint at the unreasonableness now of Mr Willcox's effort at stretching out the learning-curve in this way. 211. A relatively minor increase in one year of 7,000 M/T's in maximum capacity does not strike us as a good reason for extending the learning-curve. 212. For the sake of fairness and consistency with Mr Willcox's earlier learning-curve, 7,000 M/T's per annum should be added to the steps of his earlier learning-curve, with the result that SFI's learning-curve under the 1982 installation becomes as follow:-
213. Amended in that way, we accept Mr Willcox's learning-curve in preference to Mr Medley's in relation to SFI's capacity under the 1982 installation. 214. Put forward in support of the SFI "take-off" hypothesis was the evidence of Mr Ho, SFI's maintenance engineer and the man who installed the concast, that there was no point in increasing production before yields were improved, since, below break-even point, the more you produced the more you lost. 215. Of course, what he said about the effect of production below the break-even point is unassailably logical. If one follows that logic through, though, it would presumably have been better for SFI to produce even less than it did in the Scheme-World. 216. We feel there probably was an element of trying to make a virtue out of necessity in Mr Ho's evidence relating to SFI's stagnant production in the years preceding the "shadow". 217. Even if it is correct that SFI deliberately curtailed production until yields reached a level which could be profitable, we do not think the necessary corollary from that is SFI had the ability to increase its volume at the rate required by Mr Medley's learning-curve (which, incidentally, does not curve but "kicks up", to use Mr Willcox's words). 218. According to Mr Ho's testimony, SFI set out to improve its yields and production volume, and it succeeded on both. We think that is right, so long as success is judged in terms of likely actual production from the 1982 installation of 100,000 M/T's. On the view we take, SFI, had it not been for the threat of resumption, would actually have started producing at the rate of 100,000 M/Ts per annum in 1984/5, which is some 12 to 13 years after Mr L.Y. Leung informed Mr Y.T. Cheng, in 1972, that he was going to double the then-capacity of 50,000 M/T's per annum. 219. An unfortunate side effect of the working of litigation's adversarial approach has been that SFI implicitly gets portrayed as a failure on both the government's and its own approach on a finding from the Tribunal that probable maximum production was 100,000 M/T's, rather than 110,000 M/T's under the 1982 installation. Avoiding such a frame of reference, we regard SFI's attaining a capacity to actually produce 100,000 M/T's in 1984/5 as a success-story. That is what SFI set out to do, and, albeit rather slowly, it did it, in the face of a difficult switch from the simpler process of making mild steel rods to the more exacting technology of high-tensile rebar. 220. In usual adversarial style, there was exaggeration from both sides, with Mr Widdicombe, in a flight of fancy, describing SFI as "vibrant" at the time the "shadow" fell, while government tried to create the impression that SFI was all at sea with the technology required for mass producing high-tensile re-bars. The truth, in our view, lay somewhere between those two extreme positions. 221. Indicia of ripeness for take-off at the time the "shadow" fell, according to SFI, were discernible in:
222. Whether, at the time the "shadow" of resumption. fell, SFI had gained mastery over the continuous casting machine to the point where it could produce billets at the rate necessary for Mr Medley's version of production build-up, is pivotal to determining the significance of three of the other indicators of readiness for take-off - experienced work-force, experienced management, and improved yields - relied on by SFI. 223. Experienced the workforce might have been ; in making mild steel bars, but the indicators, by the time the "shadow" fell, if anything, point in the direction of SFI's workforce lacking the degree of experience needed for making high-tensile rebars of the right quality and quantity to meet the levels set in Mr Medley's model. 224. The main problems were, we think, getting the chemistry of high-tensile steel making right, and attaining proficiency with the continuous casting machine. 225. Employing an outside specialist in operating continuous casting machines, Mr Chui Ping-ton, some time in the second half of calendar year 1981, probably helped SFI sustain the dramatic improvements achieved in its melting yield at about that time. 226. Other factors which might well have contributed to the improved yield from the continuous casting machine in 1981 were its modification by adding auto-level controls which reduced the scope for human error, and the modification by Mr Beck of Demag, a German company which makes continuous casting machine, of the moulds' water cooling system, by removing the foot-rolls and adding more sprays. He also made it easier for SFI's workforce to control breakouts by isolating one strand from the other, so that one strand could continue to function, despite a breakout on the other. As the result of Mr Beck's work, the strand with the breakout could be relatively quickly replaced by what was described as a "cassette'', so that down-time should have been reduced. 227. Although Mr Roy Leung sought to give the impression his workforce was well trained, we do not think that to have been the case. There was not the sort of systematic training of SFI's work-force described in some of the technical journals or conference material put in evidence. SFI's workforce, certainly in the lower echelons, was left to pick up what it could on the job. People at the more senior level would go on overseas study tours of well-regarded mini-mills, but there are obvious limitations to what can be learnt that way. For example, after SFI's senior staff, on a study tour of Japan, had stood around for several days watching Toshin Steel Ltd's super-efficient performance, they were nowhere near being able to duplicate it themselves on returning to Hong Kong. 228. Management at Shun Fung during the Second Phase of the Expansion Programme preceding the fall of the "shadow" was undoubtedly enthusiastic and energetic, but rather out of its depth when it came actually to harnessing the technology of high-tensile steel rebar-making. 229. Mr Roy Leung, who was in charge of the Second Phase of the Expansion Programme and getting the new machinery to produce, lacked the necessary engineering and organizational skills for this. Although he had qualified engineers such as Mr Ho working for him, they must have lacked the specialised knowledge necessary to make a mini-mill run at a profitable level of production of high-tensile rebars. we learnt from Mr Len Leung that as late as Financial Year 1981/2, SFI was still "groping with the machinery". 230. The spheres of operation where management was shown at its weakest were in getting to grips with the problems of the continuous casting machine within a reasonable time, and in ensuring that scrap of the right quality, with the right mix, at a reasonable price, was available for the E.A.F.'s. 231. The scrap problem was also a reflection of a larger problem of lack of effective financial control of costs in the mini-mill. 232. Every single cent spent was properly accounted for in the sense of no chicanery. There was no problem there. However, there was no one to monitor that the money was being spent effectively. The basic data needed for knowing unit consumption and cost at various stages of the manufacturing process was either not collected, or not processed in a way that would make a useful management tool. 233. It is all very well for various witnesses on SFI's behalf to pooh-pooh the importance of such paper-based tasks as record keeping or analysing, and to high-light the practical side of actually getting down to making the rebars, but the situation of a dysfunction between the two sides can make all the difference between a business being at least marginally profitable or, alternatively, loss-making. 234. As the continuous casting machine never produced at a rate exceeding 4,400 M/T's of billets in any month (See Exhibits SF88C, R31 and the "Tribunal's Table" at the end of this Section), there is scope for it to be highly conjectural, on the evidence, whether, in the No-Scheme-World, combined production of billets and ingots would have been likely ever to reach the monthly level of just under the 10,000 M/T's necessary for attaining Mr Medley's projected actual annual production level of 110,000 M/T's of good product under the 1982 installation. 235. November 1980 was the month SFI's continuous casting machine reached its zenith of 71,769 piculs (4,341 M/T's) of billets : See Exh SF88C and the "Tribunal's Table" at the end of this Section). 236. We do not know what proportion of SFI's casting was billets, and what proportion ingots during the relatively significant period from October 1981 until April 1983, because something went awry with SFI's book-keeping between those dates. 237. The high yields SFI was getting in both scrap to billets or ingots, and then to rebar from October 1981, (the first month of the hiatus in concasting records), onwards, is strongly suggestive that a high percentage of billets (probably close to or above 90%), rather than ingots, was being produced. We say that for two reasons. One is that high yields from casting points to the concaster rather than ingot-moulds, since, with ingot moulds, there is a relatively high failure rate due to such factors as the mould not being sufficiently filled. The other reason is that on the rolling side, a high yield indicates billets rather than ingots, since billets are all cut to the precise length required for a bar of a particular length and diameter, so there need be none of the wastage inherent in rolling ingots. The same length of ingot is used to roll a variety of lengths and diameters of rebars, with inevitable wastage of bar ends. 238. From April 1983, when SFI's accounts again start showing the proportion of casting of ingots/billets, SFI was concasting to the extent of 95% or better most months until August 1986 when all production ceased at Junk Bay. Some months there was even 100% concasting, but of rather modest volumes around, say, the 25,000 picul (1,512 M/T's) level. 239. Despite the arrival of Mr Chui Ping-tong, the concasting specialist in mid-to-late 1981, and the modification for auto-level controls on the continuous casting machine at about the same time, we regard it as unlikely that SFI could have quickened its pace of work to the level required by Mr Medley's learning-curve, and there was nothing we gathered from the evidence about the continuous casting machine to incline us to the view that SFI's directors were wrong in their original estimate of 100,000 M/T's as SFI's likely maximum output from the 1982 installation in the No-Scheme-World. 240. An increase of the level of sales in 1981/2 is the last of the five indicators SFI relied on to show it was on the verge of take-off when the bad news of possible resumption knocked it off course. SFI had had sales at a comparable level in the first half of calendar year 1978, without any lift-off in production following as a consequence. 241. For SFI, it was contended that, superimposing a graph of SFI's 1982 sales on a graph showing Mr Medley's projected build-up to 86,000 M/T's for 1982/3, gives rise to an inference that SFI must have expected to produce 86,000 M/T's by 1982/3, for the company was unlikely to have done something so reckless as selling goods it did not feel confident it could produce. 242. As Mr Willcox pointed out, there was insufficient evidence to show how SFI's sales contracts for 1981/2 were time-related, so there is no way of telling whether SFI was acting with un-business-like folly or otherwise. 243. There is the further point that because you have sold rebars does not necessarily mean you have to manufacture them. They can be bought for re-sale from someone else. Our findings on SFI's production capacity in the light of the foregoing can be summarised as follows:
Section II. SCRAP COST 244. As the cost of scrap is the biggest single outgoing for a mini-mill - accounting, on average, for something in the order of 40% of production costs, so we were told by Mr Gillett, SFI's steel economist - the price SFI pays for this item in the Scheme-World and the No-Scheme.-World will be one of the major determinants of its profitability. 245. In some respects, SFI was better placed to acquire cheap scrap than any other buyer in Hong Kong. 246. SFI had a wealth of experience behind it, both in buying and selling scrap. From the 1950's right through till 1979, SFI had bought large ships for breaking by its own workforce, using some of the metal thus recovered for its own melting operations, and the rest for selling off to re-rollers and scrap dealers. 247. At its Junk Bay plant, it had the advantage of a scrap yard which was massive by Hong Kong standards, comprising an area over 200,000 square feet. With space at a premium everywhere in Hong Kong, a large scrap-yard conferred a considerable advantage on anyone buying or selling scrap in Hong Kong, since it reduced the occasions when, due to lack of space, one might have to forego opportunities of buying scrap at advantageous prices. 248. Of all the scrap-yards in Hong Kong, we are prepared to accept that, at Junk Bay, SFI had the best-equipped when it came to coping with outsize pieces of scrap or large chunks of broken ship. With its "Best" brand shear, (installed as recently as 1980), together with its "alligator" shear, and other miscellaneous items of cutting equipment, SFI was in an ideal position to reduce pieces of scrap to whatever size best suited its operation. 249. A suggestion was made by government that Junk Bay was too inaccessible - due partly to distance and partly to the state of the road leading there - for it to be worth the while of small scrap dealers of the "Steptoe" variety to bring scrap to SFI for purchase. According to government, it was necessary to have a collecting yard in a developed area like Kwun Tong to attract small dealers. We do not think this contention from the government has any merit. We accept SFI's evidence that its Junk Bay yard was only twenty minutes' driving time away from Kwun Tong via a road in sufficiently good condition for it to be no deterrent to any scrap dealer wanting to dispose of his wares. 250. Once at SFI's yard at Junk Bay, a scrap dealer would not be kept waiting, since there was so much space for dumping his load. This contrasts with the position at SWS's scrap-collection-yard at Kwun Tong, where, as Mr Len Leung described the position, lorries had to line up for considerable periods, each awaiting its turn to dump scrap into the barges which took the stuff to SWS's mini-mill at junk Bay. As SWS had no road to its plant at Junk' Bay, but had sea-frontage there, barging was the sole mode for SWS to receive scrap deliveries. 251. Unlike.SWS, SFI had the best of both worlds - sea and road access. 252. Although SFI was so well endowed with these physical amenities such as its large, well located and easily accessible scrap yard, plus its excellent cutting equipment, there were, however, two adverse factors of a more intangible nature which far outweighed SFI's physical advantages when it came to the point of how cheaply SFI could acquire scrap. 253. Firstly, from apparently as early as Financial Year 1975/6 even, and, certainly, during financial years 1980/81 and onwards, SFI suffered from chronic illiquidity. That emerges from the analysis by Mr Meocre Li, the government's expert witness on financial and accounting matters, in his document 42/05B at page 110 et seq., and page 122, which, by showing SFI's "Quick Ratio'' from 1974/5 onwards, highlights SFI's shortage of cash in the Scheme-World. 254. For the convenience of the reader, we repeat what we said earlier in this judgment about the "'Quick Ratio". 255. The "Quick Ratio" is a liquidity index, showing the ratio between net "quick" current assets (working capital excluding stock) and current liabilities. Stock is excluded from current assets in this measurement because investment in stock must normally be maintained to permit a company to operate. After deducting the stock from the current assets, one then divides that by the current liabilities to arrive at this "Quick Ratio'', which measures more immediate solvency. 256. We accepted Mr Li's evidence that, applying a rule of thumb, SFI's "Quick Ratio" should have been 1 or better. 257. Instead, the "Quick Ratio" over the period 1975 to 1986 was as follows :- KEY FINANCIAL RATIOS
258. This continuing shortage of ready cash forced SFI to have recourse to credit for the purchase of scrap. 259. There was no real dispute that, in Hong Kong, amongst the cheapest sources of scrap for a mini-mill, is that coming from the primary collectors (less pompously referred to, throughout the case, as "Steptoes") who go out to such places as building-sites and factories to collect "raw" scrap which, generally, will not have been graded, cut down to size, or bundled. 260. As Mr Len Leung told the Tribunal, "Steptoes do not give credit''. Bearing in mind that, on average, for the Financial Year 1980/81, SFI was getting approximately 2.8 months' credit on its scrap purchases, one quickly gathers that SFI could not shop in the ordinary "Steptoe" market that year. Neither could it in any of the four subsequent financial years (i.e. 1981/2 - 1984/5), in each of which it kept its scrap creditors waiting, on average, at least four months for payment. For 1985/6, the credit period was just over two months. 261. Thus, SFI found itself having to buy its scrap at prices higher than those on offer from Steptoes. 262. To take Financial Year 1981/2, (the last "normal" year), as an example, only 4% approximately of SFI's purchases of scrap were on a cash basis. 263. In due course, we will have more to say about the hierarchy of scrap sellers, and their prices at different stages for different grades of scrap, but, for our present purpose, which is to point out generally that SFI was strapped for cash at all material times, there is no to need elaborate. 264. The second major adverse factor affecting the price SFI paid for its scrap sprang from the difficulty SFI experienced in attaining the level of technological competence necessary for consistently manufacturing from reasonably priced scrap a sufficient volume of high-tensile rebars meeting B.S. 4449 to reach its targeted capacity. 265. From Mr Len Leung, the Tribunal learnt how, in Financial Year 1981/2, the last "normal" year before the "shadow" started taking effect, and some three years after the concaster was installed, SFI was still "groping with the machinery". The meltshop was blaming its poor results on the indifferent quality of scrap Mr Len Leung's Auto-Marine Department was providing for charging the E.A.F.s. In an endeavour to show that his department was not to blame, and that the fault lay with the meltshop, Mr Len Leung started purchasing for the meltshop scrap, which was purer and of higher density than previously, by virtue of its having undergone more processing. Unsurprisingly, Mr Len Leung had to pay more for this better quality scrap. 266. Sorting out either of the single issues (1) whether SFI's long-standing liquidity problems could be absolved from blame for the high price SFI paid for scrap under the "shadow", or (2) whether SFI had mastered. the technique of making B.S.4449 rebar from reasonably priced scrap before the "shadow", is difficult enough on its own, but when those two issues coalesce, as the government indicates they do here, the difficulty becomes compounded. 267. In practical terms, the dispute between the parties over SFI's scrap cost hinges on whether the surrogate to be selected for SFI's scrap cost during the "shadow" should be Hong Kong's only other mini-mill (i.e. SWS), or KYM Metals Ltd ("KYM"), a company specialising in the buying and selling of scrap. Which of SWS and KYM gets selected as simulacrum for the "shadow" period also bears on the scrap price to be included in SFI's No-Scheme-World accounts for 1986/7 and 1987/8, i.e. the two Financial Years immediately following resumption, and on the scrap "trend price" for 1988/9 and beyond. For the last "normal" year - Financial Year 1981/2 - there is no dispute : $475, being SFI's actual scrap price for that year, is to be used. 1982/3 - 1985/6 268. SFI's actual prices per Metric ton o?scrap in those years were as follows:-
269. According to the government's case, those very figures should be taken for the No-Scheme-World, too, during that period, on the basis that they were not in fact impacted by the threat of resumption. 270. Each side put forth its version of what, in effect, the "normalised" price of scrap should be for that period. 271. Mr Gillett, SFI's expert steel economist, introduced the concept of a "normalised" price for both scrap and rebar. 272. It was urged upon the Tribunal by Mr Gillett that the price paid by Hong Kong's other mini-mill, SWS, for the period under consideration, should be taken as the "normalised" price for SFI, the idea being that, if SFI had not suffered from the effects of the "shadow", that was the sort of price SFI might normally have been expected to pay in the No-Scheme-World. 273. Those SWS scrap figures are as follows:
274. To get a fuller picture, one needs to compare the SFI and SWS scrap prices for 1980/1 and 1981/2, as well. A detail to be noted is that the SWS scrap purchase figures, made available to the Tribunal, run from August 1980 when SWS started melting operations. (The full SWS scrap purchasing figures, on a monthly basis, are to be found in Exh SFI 211 at p.49, and, yearly, at p.87.) 275. We now set out the figures for SWS's and SFI's average cost oscrap for the Financial Years 1980/1 and 1981/2 :
276. A fundamental argument was relied on by the government against allowing Mr Gillett's "normalised" figures to be adopted for SFI in the No-Scheme-World, namely, they lacked normalcy in so far as SFI was concerned. For the government, it was contended that, unlike SFI in 1980, SWS at that time was a greenhorn when it came to the purchase of scrap, and, until about 1984/5, was following a learning-curve for scrap-purchasing in the Hong Kong market. 277. The bench-mark to be applied, according to the government, for the purpose of determining the price SFI should have paid for its scrap in the No-Scheme-World, was the price KYM had been paying for its scrap purchases at the material times. KYM has been buying scrap in Hong Kong for at least as long as SFI. Average prices for cash-collection paid by KYM for scrap in the period now being considered were :
278. In respect of Financial Year 1981/2, the Tribunal has KYM's figures only for the six months January to June 1982. 279. From SFI, there is a challenge to the admissibility of all the material purporting to come from KYM, on the basis it was mere hearsay which SFI has had no opportunity to challenge, as neither side called any witness from KYM. We overrule that objection for reasons stated at the end of this Section. 280. Another argument mounted against the KYM figures by SFI is that, unlike those from SWS, they do not amount to a continuous series starting from before the threat of resumption arose. That, in our view, is a point going to the weight to be attached to them. Obviously, they would be that bit better if they did run from July 1981 or even earlier, but we do not regard this defect as of any great consequence. 281. To facilitate understanding, we now set out a table, based on Mr Li's Report 42/05B at page 133, showing the prices paid per metric ton at various times by SFI, SWS,, and KYM, for scrap, and the relationship of those prices to each other.
(For 88/9 KYM's price is based on 10 months and SWS's on 11 months. See SF211 at page 87) 282. Of pivotal importance are the figures for 1981/2. SFI wants the Tribunal to hold that its price of $475 and SWS's price of $459, which are of a similar order, were both "normal" for a Hong Kong mini-mill that year, whereas KYM's price of $379, which was approximately 25% less than SFI's, was abnormal in the context of considering what a Hong Kong mini-mill might reasonably have expected to pay for its scrap in that year. 283. Once SFI can establish it was paying no more than a. "normal" price for its scrap in 1981/2 - the last year unaffected by the "shadow" - it then has a launching pad for a line of argument to the effect that, whereas shadowless SWS continued paying a "normal" price for its scrap in the years subsequent to 1981/2, SFI, on the other hand, suffering from the effects of.the "shadow", started paying more and more, relative to SWS, as the years went by. 284. Against that.line of argument, government contends that the correct point of departure for fixing the price SFI should have paid for its scrap, in the No-Scheme-World, is the price KYM was, on average, paying in 1981/2. (i.e. for the six months January to June 1982 on which the Tribunal has evidence). 285. From the table of relative prices presented four paragraphs back, one sees how SWS's and KYM's prices steadily converge over the years (with a slight hiccup in 1983/4), until they are more or less level-pegging in 1987/8 and 1988/9. 286. On the government argument, then, "normalcy", for scrap purchasing, was typified by KYM in 1981/2 and the succeeding years. According to the government, SWS got more and more "normal" until 1987/8, when it became completely "normal". By contrast, so the government's argument goes, SFI was already abnormal from a scrap purchasing point of view in 1981/2, and continued along that same path of paying roughly 25% (with a dip to 16% in 1983/4) more for its scrap than KYM until resumption in 1986. 287. Not only does the government seek to be exonorated from causing SFI to pay a higher price for its scrap during the "shadow" period, but also asks the Tribunal to extrapolate the premium SFI was actually paying for its scrap in the Scheme-World so that, for the post-resumption years 1986/7 and 1987/8, SFI should be treated as still paying a premium for its scrap in the No-Scheme-World, the suggested amount being approximately 20% above KYM's prices for the years 1986/7 and 1987/8. On that basis, SFI's scrap would be priced at $550 in 1986/7 and $650 in 1987/8. 288. $650 in 1987/8 constant dollars should also be the "trend price" for 1988/9, and beyond, according to the government, which argued that scrap is a production cost like any other. 289. Whilst it was common ground that all other production costs for 1987/8 in the No-Scheme-World should be treated as "trend" figures priced in 1987/8 constant dollars for the No-Scheme-World accounts for 1988/9 and beyond, SFI argued that a different approach should be adopted for the purpose of ascertaining the scrap cost "trend" figure for 1988/9 and beyond. 290. According to SFI, the scrap "trend" figures for 1988/9 and beyond need not be the same as the 1987/8 scrap cost figure. In SFI'S version of the 1987/8 Profit and Loss Account in the No-Scheme-World, the scrap price that year is shown as $546, hereas the projected "trend price" for 1988/9 and beyond is $514 in inflation-proofed 1987/8 dollars. 291. We will elaborate on the scrap "trend price" for 1988/9 and beyond when we deal with that 1988/9 period and beyond, later in this Section. 292. It is now convenient to scrutinize in more detail what SFI's actual purchase price of $475 per metric ton of scrap represented in 1981/2, the last year unaffected by the "shadow". 293. At this point, mention needs be made of the significance of the burden of proof. Unless SFI can show that, in the critical Financial Year 1981/2, it probably was buying its scrap in more or less the same way as SWS, then, leaving aside the possibility of an injection of fresh working capital, SFI cannot expect the Tribunal to regard SWS's scrap purchasing prices in subsequent years as in the nature of an index which SFI is entitled to follow. 294. Uncontroverted evidence shows how, during the Financial Year 1981/2, SFI bought all its scrap, totalling 40,148 metric tons, from nineteen suppliers. Roughly 80% of that volume came from five of those nineteen. Nearly all.of SFI's purchases that year were on credit, a mere 4% or thereabouts being for cash. The whole volume of SFI's scrap purchases that year was covered by two hundred and thirty-three invoices. 295. By contrast, SWS's purchase of some 180,000 metric tons of scrap for that same Financial Year gave rise to a figure of the order of one thousand five hundred invoices per month. Mr Len Leung agreed that SWS's purchases would generally have been for cash, in the same way that KYM's purchases were. 296. The inference is inescapable that SWS engaged in a large number of transactions for relatively small loads of scrap, whilst SFI had few transactions, but for relatively large quantities. 297. That pattern of trading has to be considered in the light of what Mr Gillett described as "The Hierarchy of Scrap Prices in Hong Kong", in his Report 31/02 at p.15. While that document emphasizes how the price rises each time scrap gets handled along the dealer-chain, a generally complementary chart (Exh SF 116) from Mr Len Leung describes the persons buying and selling at each stage. Both documents are inevitably oversimplifications of what is obviously a highly fluid situation of many different types of buyers and sellers of a lot of different categories of scrap. 298. The basic picture that emerges is of a hierarchy, with primary collectors at the bottom going out to such places as demolition sites, factories and shipyards to collect scrap. For illustration purposes, Mr Gillett showed prices on the basis of a transaction in September/October 1988. These primary collectors can, for convenience, be described as "Steptoes", but, under that rubric, Mr Len Leung also included "truckers" "demolition contractors" and others. 299. Mr Len Leung was not consistent in his terminology, since, at different times in his evidence, he called Kam Kee Metal, (SFI's biggest supplier of scrap from 1981 onwards) a. "Steptoe", a "trucker" and a "larger dealer". 300. The "Steptoes", in the example given by Mr Gillett, were buying "unsorted/mixed" scrap at source for about $400 per metric ton. The Steptoes, at that point, then have a choice of selling the scrap to either a Hong Kong mini-mill (more correctly, "the" Hong Kong mini-mill, SWS, in 1988), or to a dealer. 301. For what Mr Gillett described as "sorted and rough bundled (for local delivery)" scrap, made up of approximately 90% Grade 1 and 10% Grade 2, the Steptoe could expect to be paid just over $500 from either the mini-mill or the dealer, on a "Cash on Delivery" basis. 302. Where SFI lost out in 1981/2, and the subsequent years until resumption, was in lacking access to the Steptoes at Stage 1 of the hierarchy who are only willing to sell on a prompt cash basis. 303. Instead, SFI found itself limited to doing business with scrap sellers at Stage 2, referred to by Mr Len Leung o as "Scrap Dealers", who, naturally, add on a margin for profit above what they have paid the Stage I Steptoes, and, needless to say, there is a surcharge for a buyer in SFI's unfortunate position of requiring longish credit. 304. Besides having an outlet with local mini-mills, the "Scrap Dealers" at Stage 2 also have the option of moving up to Stage 3 of Mr Len Leung's chart, namely, to sell their scrap for export. 305. Whether those willing to sell scrap to SFI in 1981 until the time of resumption fell into Mr Len Leung's "Steptoes", "Scrap Dealer" or some hybrid classification, or whether they should be placed at some intermediate stage, is neither here nor there. All that really matters is what sort of price they were charging SFI for what sort of quality. 306. To get back to our analysis of what was happening with SFI, SWS and KYM on the scrap front in the critical 1981/2 period, we are satisfied that, to all practical intents and purposes, SWS was then paying and continued to pay prompt cash, like KYM, right through till mid-1989, beyond which the evidence does not go. 307. For reasons which elude us, arguments have been advanced on SFI's behalf that, perhaps, like SFI, SWS might have been buying on credit in 1981/2 and some of the subsequent years. While further muddying some already very murky waters, we fail to see how SFI gets helped by such arguments. 308. Suffice it to say that, we accept what Mr Len Leung himself said to the effect that, like KYM, SWS paid prompt cash. Such evidence as there was to the effect that, sometimes, SWS's suppliers might delay a few days before actually collecting their cash, and might wait till the end of the month before banking their cheques, we disregard as de minimis, and, moreover, of no relevance in the light of SFI's contention, which, we understand to be that, SWS was an exemplary; problem-free mini-mill like SFI would have been had it been spared the "shadow". 309. KYM described the prices it furnished to the Tribunal as being the "collection price" for scrap delivered to their yard "for prompt cash" (see 42/05B at 137-139). Those descriptions, along with'"cash collection" are about as self-explanatory as any words can be, and did not appear to pose any problem to Mr Len Leung who explained that the level on which SWS bought its scrap was the same as KYM's (See Mr Len Leung's Transcript, page 276). On a common sense basis, any submissions to the contrary from SFI's lawyers notwithstanding, we feel there cannot be the slightest doubt concerning the meaning of those words within inverted commas in the context of a case focusing on the practices of business relating to scrap. To all practical intents and purposes, we are satisfied that SWS bought its scrap for more or less immediate cash in the same way as KYM. Nothing in the evidence revealed any difference between the practices of KYMand SWS which materially worked to the advantage of SFI's case. The differences did not make any difference. 310. Attention can now be given to the quality of scrap each of SFI, SWS and KYM was buying at the material times. 311. In his evidence, Mr Len Leung told the Tribunal that the scrap SFI bought was the same grade as SWS's, namely, Grade 1, and SWS's in turn was the same as KYM's (See Mr Len Leung's Transcript pages 80 and 81). It emerged that all three of them were buying the grade known as "H.M.S. (No. 1)", but as we have already indicated SWS and KYM made their purchases at a different level of processing from SFI. The probability being that all three of them were buying the same grade all the way through, the inference follows that the government is probably correct in its assertion that KYM represented the "normal" market price for scrap in 1981/2, while SWS was working its way towards "normalcy" with each passing year of experience, and SFI was paying an abnormal amount in 1981/2, and continued to do so until resumption. 312. For SFI, it was argued that the scrap which KYM described as "HMS (No. 1)", and for which KYM paid $379 per metric ton in 1981/2, was not necessarily as of good a quality as the "No. 1 H.M.S." for which SFI paid $475 per metric ton that year. (We find that "HMS (No.1)" and "No.1 HMS" are one and the same thing.) 313. The Tribunal's attention was drawn to exhibit SF117, produced by Mr Len Leung, which showed that, according to the system of scrap-grading in Hong Kong, "No.1 HMS" included "Black sheet industrial pressing". We do not think there is any dispute that "Black sheet industrial pressing" is of better quality and more expensive than common or garden "No.l HMS". 314. KYM describe "Black Sheet Industrial Pressing" as "A-Press (New Cuttings)". They are one and the same thing. 315. This same stuff is also known as "Black sheet stamping". For the sake of simplicity, we will henceforth refer to Black Sheet Industrial Pressing, alias Black Sheet Stamping, alias A-Press (New Cuttings) as, "B.S.S.". 316. From Report A2/05B at page.138 it looks as if KYM probably segregated B.S.S. from HMS (No.1), selling the former at a higher price. However, it looks, too, that KYM also segregated the lower grade and lower priced "B-Press (NO.2)" from its HMS (No.1). 317. It was argued on behalf of SFI that the apparent absence of B.S.S. from KYM's HMS (NO.1) was a possible explanation of why KYM's HMS (NO.1) was so much cheaper than SFI's which embraces B.S.S. 318. According to Mr Gillett, Mr Len Leung probably knows more about scrap in Hong Kong than any other person. That being so, the Tribunal finds it of considerable significance that he regarded the grades covered by the SWS and, the KYM prices as the same (See Mr Len Leung's Transcript, page 80, line 6 and p.276). Presumably, Mr, Len Leung saw nothing untoward in the comparison of SWS's and KYM's prices for HMS No.l when he produced SF118, for otherwise he would have. remarked on it. If it had been Mr Len Leung in the witness-box who started querying whether SFI!s, SWS's and KYM's grades were the same, we would have felt constrained to ponder his evidence most carefully, but when the point is made by way of submission. only, we do not feel inclined to treat it with the same respect. If this proposition, advanced on behalf of SFI, embodied compelling logic, we would be quick to accept.it, but our attitude is rather different towards lawyerly speculations. 319. We see no good reason to depart from Mr Len Leung's contention that SFI, SWS and KYM were all buying the same grade, albeit at different levels of processing, and we regard SFI's submissions centring on SF211 at pages 93 to 95 as no better than a red herring. 320. From Exh SFI 211 at page 93, one sees that SFI, and its chosen surrogate, SWS, did buy B.S.S. separately from HMS No.l. That occurred in November and December 1985, and, presumably, the situation would have been similar at other times. The Tribunal was not informed, though, what proportion of the HMS No.l bought by SWS was B.S.S., thus giving rise to yet another grey area on the extent to which SFI's purchase of scrap compares with SWS's. In the Financial Year 1985/6, SWS's average price for scrap was within a mere 3% of KYM's price, which suggests the quality of SWS's was little different from KYM's. As likely as not, the position was the same in earlier years. 321. With B.S.S., in our view, eliminated as a likely explanation of why SFI's average scrap price was so much dearer than KYM's, we turn to consider whether the evidence suggests some other feature to explain the price differential between SFI and KYM. 322. We know that Mr Len Leung claims to have bought scrap in the Financial Year 1981/2 of a purity and density higher than in the past. 323. How much higher that density and purity were cannot be ascertained in any objective way, since SFI destroyed all its operational records, only those for 2nd - 4th January 1982 having survived (Exh SFI 81). 324. We are thus left just with Mr Len Leung's word on this. 325. Just how wide was the variety of scrap SFI used can be gleaned from those few extant records (Exh SFI 81) which, inter alia, referred to "Runner scrap" "Cast from scrap mould", "Black sheet mild" "Loose black sheet mild" "sheared scrap" "1st Choice Grade Scrap", "shipbreaking scrap" "Mild steel bar end", "Loose first choice scrap" "Loose scrap from 40 gallon steel drum", ". stampings (Grade 1) material", and so on. 326. Were that sort of detail available over a prolonged period, the steel experts, Mr Medley and Mr Willcox, should have been in the position of eing able to tell the Tribunal how high and pure the density and quality of SFI's scrap, in fact, was. 327. According to Mr Len Leung, the extra-dense and extra-pure scrap he was buying in 1981/2 came solely from Kam Kee Metal, ("Kam Kee"), which, from that time until resumption, was SFI's biggest single supplier of scrap. A strange feature of the evidence is that, from the Journal vouchers of SFI which Mr Mebcre Li saw, the scrap from Kam Kee was no dearer than from the other suppliers to SFI at the same time. 328. Moreover, it turned out that even after 1981/2, the time Mr Len Leung said he was experimenting with the denser, purer scrap, Mr Len Leung continued to buy this denser, purer scrap, although SFI no longer needed scrap of such high quality, according to Mr Len Leung. 329. Mr Len Leung's explanation of why, after 1981/2, he continued to buy better quality scrap than he needed from Kam Kee was that Kam Kee was one of the few suppliers still willing to give SFI credit, and SFI, in effect, had to take whatever Kam Kee offered. 330. The whole situation concerning SFI's purchase of scrap in 1981/2 and beyond is strange and confused, and becomes even more so when one tries to factor in the elements of sorting and credit. 331. There was no dispute that at least in 1981/2, SFI purchased some sorted material from Kam Kee. That was why that material was more pure and dense than usual. 332. Mr Tsang, the Chairman of the Hong Kong Metal Merchants Association, and managing director of KYM, explained; in a letter faxed on 23rd December 1988 to the government's accountants, Arthur Andersen and Company (Exh R42/05B at page 137), how his company estimated the cost of sorting scrap purchased on a cash collection basis (in effect from Steptoes) at its yard. I now set out the contents of that letter :-
when those ex-warehouse figures for the cost of sorting were put to Mr Len Leung, he described them as reasonable (see Mr Len Leung's Transcript at page 204), and went on to point out that a margin should be added on for the dealer's profit. 333. There was no suggestion that SFI made any purchase from KYM, and the letter just set out is only significant for establishing the principle that the cost of sorting gets passed along the chain to the ultimate purchaser. 334. How much of the material SFI purchased in 1981/2 was sorted, and how much SFI paid above the "Steptoe" primary collection price for such processing, is a grey area in the case. The uncertain nature of the evidence must inevitably work to SFI's disadvantage, since it carries the burden of proof of what was probable. 335. Another clouded area is the interest element SFI paid its suppliers of scrap on credit. 336. We do not believe Mr Len Leung when he said in re-examination it was only about $10 to $20 per long ton in 1981/2 (A long ton is 16.5 piculs, whilst a metric ton is 16.5345 piculs.) 337. When, on the first day of his cross-examination, it was put to Mr Len Leung that SFI was paying 25% more than KYM did for its scrap in 1981/2, he said he knew SFI had paid a high price, but did not know the percentage. When he used the words "a high price", we certainly did not gain the impression, then, that Mr Len Leung was seeking to convey the idea that SFI was paying some trivial amount like $10 to $20 extra, per metric ton, for credit. To Mr Len Leung, $2 to $3 difference per picul between buyers of scrap was of no significance, being nothing more than ordinary variability. That $2 to $3 per picul works out roughly as $32 and $48 per metric ton, respectively. 338. When making his observation about "a high price" for credit, we gathered, in context, that he had a figure in mind more of the order of the 25% (which would be over $100 per metric ton) the government was suggesting, rather than the 3% to 4% he contended for under re-examination. While it is correct that he did not actually accept the government's suggestion of paying 25% for credit, it was, we think, significant that his immediate reaction was not to reject 25% as preposterous. One would then have expected him to laugh 25% out of court, rather than say he knew SFI paid "a high price". 339. Having put his foot in it on the first day of cross-examination, he thereafter tried to extricate himself, and, in so doing, lapsed into the untruthful. 340. The impression we got was that, in unguarded moments under cross-examination, Mr Len Leung let the cat out of the bag over the difficulties SFI was still experiencing as late as 1981/2 to get the technology of high-tensile rebar-making right, and how the tight cash-flow left it vulnerable to paying whatever price was demanded for scrap by the handful of credit-granting dealers still willing to do business with it. 341. From Mr Gillett, we learnt that the scrap business is a tough trade. We do not suppose that will come as a surprise to many people. 342. On the government's hypothesis that KYM represented the market price of scrap, the effective interest rates SFI must have been paying were savage. On the simplifying assumption that SFI enjoyed four months' credit on its scrap purchases in each of the Financial Years 1981/2 - 1987/8, the effective interest rate it paid, when contrasted with KYM's cash purchases in those years, was a minimum of 47.83% and a maximum of 79.23%. We set out the table from Exhibit R82 illustrating this :- Shun Fung Ironworks, Limited
Notes: N.B. This schedule is prepared on the same basis as page 92 of SF211, with the following amendments: (a) Kwong Yick prices are used as the basis for the comparison with Shun Fung since these represent a consistent series; (b) a credit period of 4 months is taken instead of 3 months, since this is consistent with the DHS calculation of the period of'credit taken by Shun Fung from its scrap creditors in the calendar year immediately preceding the shadow of resumption (SF214)." 343. By contrast, SFI's contention that it should be compared with SWS on the price of scrap was illustrated by SF 211 at p.92A, which shows the interest rate position on the basis of SFI's actual average credit periods from 1981/2 - 1985/6, and a projected period of three months credit for each of the years 1986/7 and 1987/8. We now reproduce SF211 at p.92A : Shun Fung Ironworks, Limted
Notes: 344. From SF211 at p.92A, it is apparent that the effective interest rates start tending towards the astronomic from 1984/5 and onwards. On behalf of SFI, it was argued that, the relatively low effective interest rates disclosed by SF211 at p.92A for the Financial Years 1981/2 to 1983/4, are inherently more likely than the sky-high rates for those same years in the government's exhibit, R82. The high rates for 1984/5 and 1985/6 can be explained away, so SFI argues, on the "shadow" doing its work by then. 345. Divorced from the rest of the evidence in the case, any reasonable man would probably conclude that the lower effective interest rates for 1981/2 to 1983/4, revealed by SF211 at p.92A, do look intrinsically more likely than the equivalent figures in Exh. R82. 346. However, it is not permissible to look at those figures in isolation : one must see them in the context of all the evidence in the case. 347. The over-all evidence in the case includes factors, such as SFI's probably getting purer scrap which has undergone more processing than that bought by either SWS or KYM, and the circumstance that SFI was asking for credit in what is normally a cash business. Mr Meocre Li-likened SFI having to ask for credit in this line of business to someone who is in the unhappy position of needing to borrow from loan-sharks. We regard the comparison as perhaps going slightly too far, but we do incline to the view that dealing in scrap iron is unlikely to be one of the more genteel areas of business, and rates of interest for credit bordering on the extortionate leave us unsurprised. 348. The differences between what SFI, SWS and KYM each paid for its scrap depend on a miscellany.of factors besides credit, and, in context, we do not consider that Exh. SFI211 at p.92A, nor Exh. R82 assists SFI. 349. Neither do we think that one should overlook the circumstance that when Mr Len Leung gave his evidence, it was obvious that the rates of interest to which he referred were not time-related, and there was nothing in the evidence to suggest that either Mr Len Leung or Mr Roy Leung thought in terms of effective rates of interest. For example, towards the foot of p.203 of Mr Len Leung's transcript, Mr Carnwath for the government asked Mr Len Leung in relation to scrap dealers, what sort of margin he would have had to pay. Mr Len Leung's reply at p.204, line 4 was "well Mr Carnwath has put it that it would be about 25 per cent and that I do agree". Mr Len Leung was simply saying by what percentage one figure exceeded another. 350. Poor cash flow explains why SFI was unable to buy any more ships for breaking after 1979, and was certainly a major cause of SFI consistently having to pay hefty premiums above what KYM paid for its scrap in 1981/2 and beyond. 351. An obvious cure for SFI's chronic illiquidity would have been for SFI's parent, New World, to have advanced sufficient funds to enable SFI, like SWS and KYM, to pay cash for its scrap. 352. It was agreed that, if in the No-Scheme-World, SFI were to adopt a policy of buying its scrap for cash, it would require as working capital for all purposes the amounts of money listed in column "A" below. The working capital necessary to carry on with buying scrap on credit is listed under column "B".
353. Those sums set out above fund much more than just scrap: they also fund, for example, inventory and debtors. We accept the assertion from SFI that, all that was needed, by way of funds to finance the limited objective of switching from credit to cash for its scrap purchases, was the cost of one month's supply of scrap, that being the approximate volume of scrap kept by SFI in its scrap-yard at Junk Bay. We gathered from Mr Meocre Li that he was of the same mind on that. 354. If the 16 - 26% difference between KYM's cash collection price and SFI's actual price for scrap in the Scheme-World from 1981/2 to 1985/6 was explicable on the basis of a premium being paid by SFI for credit, then, according to Mr Stewart Leung, the director of New World responsible for overseeing New World's subsidiaries, New World would have made available to SFI the necessary working capital for it to switch from credit to cash in the No-Scheme-World, provided this made commercial sense. 355. Mr Stewart Leung was sure that no reasonable businessman would be willing to pay effective interest rates of the magnitude set out earlier from Exh R82. 356. More weight could be accorded to Mr Stewart Leung's claim that New World would have come to SFI's rescue in the No-Scheme-World from the high effective interest rates disclosed by Exh. R82, were it not for the circumstance that, even on the figures SFI puts forward in SF211 at p.92A, the effective rates SFI was paying in 1984/5 and 1985/6 were 55.37% and 135.61%, respectively. 357. The high rates for those two years can be explained away on the basis that, by then, the effects of the "shadow" were intensifying, according to the argument put forward on SFI's behalf. 358. Whether one adopts the effective interest rates relied on by the government in Exh R82, or by SFI in Exh SF211 at p.92A, New World invariably chose, ultimately, to make the money available to SFI to pay its scrap creditors. 359. It makes no more sense for SFI to have paid an effective interest rate of 135.61% in 1985/6 as shown in SF211 at p.92A than to have paid the 73.53% for 1984/5, say, as shown in Exh. R82. 360. It is perfectly understandable that New World was no longer willing to advance further funds to SFI for new plant and machinery once the threat of resumption existed, but it is less easy to understand why it would have allowed SFI to buy its scrap so dearly, unless, of course, SFI failed to inform New World of the difference between buying scrap on credit rather than for cash. We think that, as likely as not, is the explanation, with the left hand in the Scheme-World, not knowing what the right hand was doing. If the Leung family in the Scheme-World, against a background of a faltering economy, started asking New World for substantial infusions of working capital that might have tilted the balance in the direction of New.World deciding to get out of the steel-making business, once and for all. 361. The Tribunal was not impressed by Mr Stewart Leung's assurances along the lines that New world had at all times been prepared to stick by SFI through thick and thin. 362. The off-stage voice of Mr Alex Chow, one of the financial controllers of New World, in a working paper of Price Waterhouse, SFI's auditors, saying, in the dark days of the autumn of 1983, it was unlikely SFI's ironworks operation would be continued, whatever the outcome of the threatened resumption, has far more of the ring of truth about it than Mr Stewart Leung's testimony. (See Exhibit R33, page 000013). Mr Roy Leung thought that probably Mr Alex Chow was responding to a request for information from Mr John Lee, the person in Price Waterhouse who looked after N.W.D.'s accounts. 363. Certainly, New World knew exactly what SFI was spending in the sense that New World counter-signed all SFI'S cheques. However, New World's staff do not know the ins-and-outs of running a mini-mill, and, in practice, would have lacked the knowledge to query whether SFI might not have bought its scrap in some cheaper way. 364. We did not accept what Mr Stewart Leung said when he tried to make out that the accounting staff in New World would check with suppliers to make sure SFI was not paying too much for its scrap. 365. The impression we got was that SFI was left alone by New World to get on with the day-to-day running of its business. Mr Stewart Leung indicated as much, and even agreed that New World would not concern itself with such day-to-day matters as whether SFI was buying its scrap on, say, three to nine months credit in 1981. 366. On behalf of SFI it was argued that, with the build-up of sales in the No-Scheme World, New World would not have hesitated over injecting the necessary working capital into SFI for it to buy scrap for cash. 367. A valid reply to that was, we felt, that in the Scheme-World, New World could have saved much by, for example, advancing funds to SFI during the Financial Year 1981/2 when SFI bought the not insubstantial total of 40,148 metric tons of scrap. New World refrained from funding the purchase of scrap for cash in the Scheme-World, and the question posed for the Tribunal is whether it is reasonable to suppose New World would have changed its practice in the No-Scheme-World, and, if so, when. 368. From 1979 onwards, SFI's cash-flow in the Scheme-World was gradually getting tighter and tighter, according to Mr Len Leung. One reason advanced, on behalf of SFI, for why New World did not furnish the funds to buy scrap for cash from 1979 onwards was that, during that time, New World was financing the acquisition of new plant and machinery for SFI. That is not by itself necessarily a convincing reason for withholding working capital from SFI. There is nothing incompatible between making money available for new equipment and for working capital at the same time. SFI wants the Tribunal to find that, in the No-Scheme-World, as soon as New World had finished financing SFI's new plant and equipment for the expansion, New World would then have started supplying SFI with more working capital. 369. New World having a policy or practice of funding only the acquisition of new capital. assets, but not working capital for day-to-day needs of its subsidiary, would not, however, strike us as unreasonable, and could well be another explanation of why New World behaved as it did over SFI's scrap purchasing in the Scheme-World. On the view we take, the same would probably have happened in the No-Scheme-World as in the Scheme-World. 370. On the evidence as a whole, we see no reason to disagree with Mr Stewart Leung's observation that, in relation to scrap, SFI did not have a policy as such : if the cash-flow was insufficient, SFI would buy on credit, but were the cash-flow to suffice, it would pay cash. He mentioned a third possibility, namely, SFI approaching New World for more working capital. 371. From the evidence we have heard, we gather the impression that SFI had led a hand-to-mouth existence from at least 1979, and probably even earlier. Even if Mr Len Leung really did think in 1981 of switching from credit to cash, which we doubt, we regard it as unlikely that SFI would have asked New World for the money required, or that New World would have been forthcoming, if asked. Based on the evidence, our view is, that, in the No-Scheme-World, it would not have been before SFI started getting positive cash-flow from its operations that it would have changed from credit to cash. 372. Prior till then, SFI's actual price for scrap in the Scheme-World would also have applied in the No-Scheme-World. 373. There is also the point that economic conditions were difficult in Hong Kong from 1981/2 through till about 1985/6, which helps incline us to the view that New World would have been reluctant to make outlays of new capital at such a time. 374. Far more likely, we think, is that SFI would have waited until its own cash-flow was, positive before starting to pay cash for its scrap. 375. In exhibit R80 Mr Li has set down his view on the factors involved in SFI's switching from credit to cash purchasing of scrap in the No-Scheme-world. Paragraph 3 of that exhibit identifies three factors to be considered, namely :-
376. Our answer to (a) is that we think SFI would have made the switch as soon as it had sufficient cash-flow to do so. That would have been in the Financial Year 1986/7. That year SFI would have had sufficient positive cash flow to buy some of its scrap requirement for cash. It would not be until Financial Year 1987/8 and thereafter that SFI could have switched over to buying all its scrap for cash. We regard it as probable that, all along, both Mr Len Leung and Mr Roy Leung would have had a good idea of the cash collection price of scrap, and would have been aware that SFI was paying a hefty premium above that for credit. 377. That premium was usually of the order of 25% according to the government's case which we have accepted on this point. 378. Bearing in mind the size of that premium, it must have been self-evident to both Mr Len Leung and Mr Roy Leung that SFI could not fail to benefit from such a change. Moreover, as such a change was to SFI's advantage even on the unfavourable assumptions made by the government in its Exhibits R162 and R169, and in Mr Li's report 42/05E 2, pages 11 - 16B-G, there can be no doubt that, a fortiori, the change would have benefitted SFI in view of our more favourable findings for SFI. 379. We do not think that a conclusion to that effect could be drawn only after performing an exercise along the lines demonstrated by the exhibits to which we have just referred in the previous paragraph. For the sake of completeness, we have had those exercises done with our figures, and they are to be found in the Tables at the end of the present Section. The Tables are copied from DHS's Bundle dated May 1992 entitled "Revised Shunde Projections". The only exception to that is the DHS equivalent to page 16B of 42/05E2, which is to be found in the DHS Bundle, dated December 1991, page 003, the column headed "Alternative 2". That column does not incorporate the amendments made necessary by the DHS Bundle, dated May 1992. 380. In performing the exercise just described of identifying up-to-date versions of pages 11 - 16B-G, we were greatly helped by the Table of Concordance (Ref B158 PL451/AJW/in/20.12.91) which accompanies the Bundle dated December 1991 of DHS Revised Appendices. 381. Perhaps in a border-line situation, a sophisticated exercise like that would be needed before arriving at a conclusion, but, in a more obvious case like the present, a practical businessman like Mr Roy Leung would, we think, have intuitively recognized the financial advantage. 382. In relation to Mr Li's factor (b), we see no reason why SFI, on the assumption of having cash to put on the table, could not have reached the same level as other cash purchasers, like SWS and KYM, almost straight away. 383. Because of SFI's long experience in buying scrap in Hong Kong, we do not think SFI would have had to undergo a long learning-curve similar to SWS's when the latter took from 1980 to 1985 for its prices to come within 5% of KYM's. 5% also happens to be the variability referred to by Mr Tsang in R42/05B at p.137 for the cash collection price of scrap in Hong Kong. That 5% can be plus or minus. It is, in effect, another way of putting Mr Len Leung's observation that a difference of $2 to 3 per picul in the prices different purchasers pay for scrap is neither here nor there. 384. Lastly, there is Mr Li's factor (c), the impact another major scrap purchaser would have on the market. 385. With SFI as another major purchaser in the No-Scheme-World, we think that all purchasers would finish up paying more. 386. We are satisfied that SFI's requirements of scrap in the No-Scheme-World would have had an inflationary effect for all purchasers, but not for the reason Mr Li relied on. 387. According to Mr Li, the reason SFI and the other cash purchasers would have had to pay more was because of SFI's transfer from the credit to the cash market. 388. In the context of available supplies of scrap in Hong Kong, we reject Mr. Li's view about there being separate cash and credit markets. 389. As we see the situation, all purchasers, of scrap in Hong Kong whether for cash or credit, would have had to pay more in the No-Scheme-World, because, from Hong Kong's same limited pool of scrap (estimated by Mr Gillett to be about 465,000 metric tons for 1985/6), SFI would have been taking increasing amounts, year by year, corresponding with SFI's increase in production. 390. In the No-Scheme-World, on our findings, SFI's requirements of scrap each year would have been as follows :-
391. Each year, SWS required approximately 180,000 metric tons of scrap. The balance, after SFI's and SWS's needs were satisfied, would be bought by local scrap merchants for export. 392. As Mr Carnwath conceded, there was some force in Mr Gillett's contention that SFI's intake of scrap would have increased slowly - not a sudden leap, but just a small, amount more each month, spread out over the four years it would have taken, on our findings, for SFI to reach its full production capacity of 110,000 M/Ts in 1985/6. 393. Despite the slow nature of SFI's increasing demand for scrap, we think the probabilities favour a gradual increase in price as time went by. 394. The numbers speak for themselves. 395. For example, in the Scheme-World in Financial Year 1984/5, SFI purchased a mere 21,405 metric tons of Hong Kong's estimated 487,000 metric tons of scrap, so that SWS, KYM and the ten or so other major dealers in scrap that year, had 465,595 metric tons to compete for (References: Mr Gillett's report, 31/01, page 22 and Exh SF 211, page 89). In the No-Scheme-World for that same year, SFI would have required approximately 113,000 metric tons, so that the others would have found themselves striving for a share of the remaining 374,000 (approx.) metric tons, instead of 465,000 (approx.) metric tons. 396. The point is made more forcefully if one lumps the mini-mills together on one side, and the scrap dealers on the other. In.the Scheme-World in 1984/5, SWS and SFI between them took approximately 210,000 metric tons of scrap; leaving 286,000 metric tons (i.e. 487,000 - 201,000) for the dealers. In the No-Scheme-World, though, for that same year, the dealers would have found themselves left with only 194,000 metric tons (i.e. 487,000 - 293,000) to fight over. 397. On basic principles of supply and demand, we think there is a likelihood of all buyers paying more as SFI took an increasing share of Hong Kong's scrap. 398. Where we disagree with Mr Li is over the impact. SFI's increasing demand would have had on the prices of all buyers, whether cash or credit, at the time when SFI was still making its purchases on credit. It would not have been SFI's switch into the cash market which somehow heralded the entry of a big new buyer, sending everyone's prices up. The inflationary effect would, we think, have started manifesting itself even when SFI started increasing the volume of its purchases on credit. 399. Logically, there is an argument for adding an inflation surcharge to the premium SFI was paying so long as it bought on credit in the No-Scheme-World, but, since the government has not asked for that, we do not propose burdening SFI with it. We cannot, however, see any good reason why, from the time when, in the No-Scheme-World, SFI changes to cash purchasing of scrap, the government should have to forego the benefit its case should derive from the Tribunal's recognition that, in the No-Scheme-World, SFI's increased buying would have had an inflationary impact on all buyers. 400. No one could deny that the all round price increase we contemplate for scrap in the No-Scheme-World will be difficult, if not impossible, to measure, but such difficulty does not permit the Tribunal to ignore the probability of such an increase. 401. There are two ways in which the Tribunal can deal with this increase. One, is to adjust the discount rate upwards for imponderables and vicissitudes when valuing SFI; the other, is to adjust the cash-flow, so as to make allowance for this contingency of a higher scrap price. We prefer the second course, as it smacks less of fudging and tinkering with the discount rate. 402. Adopting a robust approach and doing the best we can, we propose treating all buyers' prices as increased by 10% from the time SFI moves across to a cash basis for its scrap. It is no more than coincidence that Mr Li's suggested increase was 10%, and it is not because he suggested it that we have adopted it. 403. Of course, from among the previous scrap purchasers, the only ones of any interest to us are SFI's chosen surrogate, SWS, and KYM, the latter being the one the Tribunal regards as the most suitable for present purposes. 404. As in the No-Scheme-World, SFI does not enjoy any positive cash-flow and, hence, cannot change to cash purchasing of scrap until 1986/7, we only need make an upward adjustment to KYM's scrap purchase price from that year onwards. 405. Thus, in 1986/7, such scrap as SFI's cash flow permitted it to purchase for cash would have cost SFI $499 per M/T (i.e. KYM cash price plus 10%). That figure of $499 per M/T becomes the "normalized" cash price in the No-Scheme-World for 1986/7. 406. On the same basis, such scrap as SFI bought for cash in 1987/8 would have cost $595 per M/T, and in 1988/9 when SFI's cash flow would at last have permitted to buy all its scrap for cash, its price would have been $595, the figure at which in 1987/8 constant dollars scrap would remain from 1988/9 onwards. $595, therefore, becomes the "normalized" cash price in the No-Scheme-World for 1987/8 and beyond. 407. Some further points made by Mr Gillett as to why he did not think anyone's price of scrap would rise in the No-Scheme-World can.now be briefly commented on. 408. He suggested that the two mini-mills, SFI and SWS, would act as a cartel, fixing the price at the same level as if SFI were not in the market. We regard the cartel idea as far-fetched, and contrary to SWS's experience when it first started buying scrap in 1980. SWS started off paying more than SFI for scrap, and it can be seen from Exhibit SFI 211 at p.137 that the prices of SFI and SWS never ran in tandem in the Scheme-World, and, in our view, there is nothing to suggest they would do so in the No-Scheme-World. 409. Moreover, as SWS and SFI are not the only players in the market, it is difficult to see how they could operate as a cartel, even if so minded. 410. It was Mr Gillett himself who explained how mini-mills sometimes play the market by, for example, stocking up with scrap to a point where they can teach "Steptoes" a lesson by refusing to buy any of their scrap. Those are presumably the sort of games that the "Steptoes", in their turn, can play, too. A market where those sorts of free-wheeling manoeuvres can take place is, surely, inconsistent with the existence of a cartel. 411. There was, also, a "wild card" of cheap scrap from China, which Mr Gillett prayed in aid as a "counter-vailing force", for the purpose of capping any increase in SWS's (and, hence, SFI's) scrap price in the No-Scheme-World. 412. Presumably, we would have thought, the prospect of that cheap China scrap is already factored into SWS's and all the other buyers' prices in the Scheme-world. The No-Scheme-World will be different : there will be an additional major buyer, SFI, competing to buy from the same sum total of scrap (including potential scrap from China) as existed in the Scheme-World. SFI, the extra buyer, would, in our view, disturb the equilibrium of the market by increasing competition amongst buyers. Adam Smith's "hidden hand" would be there, doing its work. Accordingly, it is likely to become more of a sellers' market, and it is reasonable to expect the buyers to have to pay more. 413. Another of Mr Gillett's points was that scrap prices for SFI did not increase when SWS first started buying scrap in 1980. On the contrary, prices went down see: Exh SF211 at p.137. Nothing spectacular happened, either, when SFI started bowing out of the market from 1982/3 onwards. We do not regard that "post hoc, propter hoc" line of argument as, by itself, persuasive. There could be so many different variables operating in the market at those times that one is not in a position to say what significance should be attached to the direction of the market. 414. There remain a few miscellaneous matters concerning KYM which call for comment. 415. It was suggested on SFI's behalf that KYM's monthly figures for the January to June part of the crucially important Financial Year 1981/2, as shown in Mr Li's report 42/05B at p.135, might be an unrealistically low offer-to-purchase price, set by KYM at such a level because it did not really want to purchase-any more scrap at the time, its scrap yard then being already full. We regard that as a fanciful possibility, deserving of no weight from the Tribunal. Looking at the pattern of figures for January to June 1982, we see KYM's scrap prices rising moderately from $347 per metric ton in January to $397 in June. On behalf of SFI, our attention was drawn to the circumstance that on several occasions for two or three months at a time, KYM's cash collection price remained the same. We see no significance in this. There is nothing to support a hypothesis that KYM was trying to deter "Steptoes" from selling to it. 416. A point was taken on behalf of SFI to the effect that KYM's figures should not be treated as the market price, because of Mr Tsang's acknowledgement in his letter dated 23rd December, 1988, (Mr Li's Report 42/05B at p.137), that other scrap dealers' prices should be treated as deviating 5%, plus or minus, from KYM's. 417. We do not regard that circumstance as militating against the use of KYM's prices as the market price for cash purchases. Such variability indicates that the cash market within Hong Kong for scrap is somewhat crude when compared, for example with the Hong Kong Stock Exchange. Scrap merchants like KYM do not openly announce the price they have just paid a "Steptoe" for some scrap, which contrasts with a buyer in, say, the stock-market where the latest purchasing prices are continuously made public for all to know. 418. It is the lack of information which makes the Hong Kong scrap market, somewhat rudimentary. Both buyers and sellers are to a considerable extent operating in the dark about the price of others' deals, but that does not mean there is no market. Information will gradually get through to those operating in the scrap market as to what prices others are paying, but such information will be far from instantaneous. We think that the likely situation in the Hong Kong market for scrap is that cash buyers and sellers have a rough idea within a range, of what, at any particular time, is a reasonable amount to pay for scrap. 419. We regard it as reasonable to accept that the Hong Kong market price for scrap falls within the range of plus or minus 5% of what KYM was paying. 420. As mentioned previously, in a slightly different context, Mr Len Leung accepted the notion of prices falling within a range when he explained that a difference of $2 to $3 per picul between what one trader and another paid was normal, and of no significance. 421. Having said all we intend to say about KYM's prices, we now turn to the liminal point of whether the evidence purporting to come from KYM (i.e. the evidence embodied in Mr Li's report R42/05B at p.137 to 140) should all be disregarded as hearsay. 422. In considering whether this objection on behalf of SFI should be sustained, we have endeavoured to treat the matter in a common-sense, business-like way, asking ourselves and answering three questions, namely
423. During the course of the hearing, we learnt that Mr Len Leung knows Mr C.Y. Tsang, and there was not the slightest hint of criticism of Mr Tsang's character by Mr Len Leung or any other witness. Mr Leung agreed that Mr Tsang was an authoritative source. We see no reason to regard the material as other than truthful. 424. We regard it as no more than forensic exaggeration for SFI's counsel to have suggested that the admission of this evidence, untested by cross-examination, would amount to a breach of natural justice. All that was needed to cure that was to ask the Tribunal for a subpoena directed to Mr Tsang. 425. Another miscellaneous matter on which we wish to say a few words relates to an observation made by Mr Len Leung that only end-users, namely, mini-mills, knew the effect of B.S.S. From that, we presume he wanted the Tribunal to infer that only mini-mill operators were competent to talk authoritatively about B.S.S. A corollary to that would be that the Tribunal should attach no weight to the KYM evidence on B.S.S. 426. What Mr Len Leung says about only end users knowing the effect of B.S.S. might well be correct, but, nonetheless, we feel, irrelevant. 427. A scrap metal merchant such as KYM appears on evidence to have no difficulty recognizing B.S.S. when buying and selling the stuff. 428. Kam Kee Metal, for example, knew enough about B.S.S. to be able to trade it (see Exh SFI 211 at p.93), and we do not doubt that Mr.Tsang of KYM, as well as the rest of Hong Kong's scrap merchants, can recognize this relatively valuable commodity when they see it. 1986/7 and 1987/8 429. As already indicated, we have already found that in the No-Scheme-World, SFI would have paid $499 per metric ton for such of its scrap as it bought for cash in Financial Year 1986/7, and $595 on the same basis for 1987/8. SCRAP "TREND PRICE" FINANCIAL YEAR 1988/9 AND BEYOND 430. It is common ground that a figure in 1987/8 constant dollars (or, the same thing, using Mr Best's words in 33/01, para 125: "... 1987/8 dollar values ...") should be shown as the cost for scrap in the Loss of Profits Claim for Financial Year 1988/9 and beyond. 431. That figure has, somewhat loosely, been described throughout the hearing as, a "trend" figure. 432. In the context of the corresponding rebar price for 1988/9 and beyond, the word "trend" was frequently used in the technical sense of a "trend-line", determined by regression analysis as in Exhibit,SF 34(c). Such a "trend-line" was also sometimes called, "the line of best fit". 433. When it came to scrap prices, both sides agreed that, based on scrap prices from 1980/1 to 1987/8, there was no discernible trend, the prices throughout that period being rather volatile. 434. Although Mr Gillett caused a regression analysis to be done for those scrap prices (Exhibit SF35), neither side attempted to make serious use of it, except to point out how volatile prices had been. 435. In a somewhat half-hearted way, Mr Gillett said that, if anything, the regression analysis pointed towards a downward trend, but he did not follow that up with any particulars. On the other hand, we ourselves can see, from the graph Exh SFI 211 at page 137, that, if one looks at scrap prices over the longer period of 1978/9 to 1988/9, there does appear to be a somewhat insignificant upward trend. Although there is the authority of Mr Gillett for detecting a downward "line of best fit" by eye, we regard the data in the present instance as too indeterminate for that (see Mr. Gillett's Transcript page 299, line 18). 436. Bearing in mind how the case.was presented before us, we are going to treat the average yearly scrap prices up to 1988/9 as trendless. 437. Unburdened of the need to detect the significance of regression analysis in this context, the Tribunal can resort to more old-fashioned means of coming up with a sensible figure for 1988/9 and beyond. 438. A further point agreed on by both sides was the difficulty of predicting scrap prices. Mr Gillett opined they were even more difficult to predict than rebar prices, and he is probably right on that. 439. Despite the difficulty, the Tribunal is of the view that a rational approach has to be adopted, and does not favour simply plucking a figure out of the air. 440. As Mr Gillett remarked, there are several sensible approaches which can be followed. 441. Through the combined efforts of Mr Gillett and Mr Best, SFI initially proposed a sensible method to arrive at a scrap price to be used for 1988/9 and beyond. 442. When Mr Gillett wrote his First Report (31/01) in February 1988, he already had SWS's actual scrap prices for the first half of Financial Year 1987/8 (i.e. for July to December 1987), and he forecast a figure of $514 for the full Financial Year 1987/8, based on sensible reasons to be found at pages 25, 26, 42 and 51 of his Report 31/01. 443. Later, in that same Report, (at page 38), he used that same figure of $514 as an estimate of the price at which the proposed new mill at Shunde would buy its scrap. That was obviously a figure in 1987/8 constant dollars. In context, it was clearly a "current trend value", in the same way as the estimated rebar sale price for Shunde was (at page 37 of the report). That trend figure of $514 in 1987/8 constant dollars was also adopted by Mr Best for the Loss of Profits Accounts for 1988/9 onwards. 444. One reason given by Mr Gillett, (and adopted by Mr Best), for choosing that figure of $514 as a future trend figure for scrap was that the Finanacial Year 1987/8 was likely to be neither a very high, nor very low, year for scrap prices in Hong Kong (See 31/01 at p.38 and 33/01, paragraph 132). A similar "high", "low" approach was also at first used by Mr Gillett, followed by Mr Best, for a rebar price in 87/8 constant dollars for the Loss of Profit Accounts for 1988/9 onwards : (See Mr Gillett : 31/01 page 37, and Mr Best : 33/01, paragraphs 125-131.) 445. That Report from Mr Best (33/01), like Mr Gillett's, (31/01), was made in February 1988. 446. At page 51 of Mr Gillett's Report 31/01, in "Figure 2, Price of Scrap For SFI and SWS 1978/9 - 1987/8", Mr Gillett's graph shows what is obviously intended to be SWS's scrap price of $514 for 1987/8 as "Provisional". 447. No doubt impressed by Mr Gillett's and Mr Best's restrained and rational approach in their February 1988 Reports, Mr Li was, first of all, prepared to accept their estimate of $514 in 1987/8 value dollars for Loss of Profit Accounts for 88/9 and beyond. 448. Common sense would dictate that, if time showed Mr Gillett to have been wrong, (either too much or too little), in his forecast of $514 as SWS's (and, hence, SFI's) scrap price for 1987/8, he, (followed by Mr Best), would substitute the actual figure in the Loss of Profits Accounts for 1988/9 and beyond. 449. SWS's actual scrap,price for 87/8 turned out to be higher than Mr Gillett's forecast, the actual figure being $546. 450. It also turned out that he was wrong in his forecast of the rebar price for 1987/8. He had forecast too high a figure. Having predicted $2,536 per metric ton, the actual turned out to be $2,458., The rebar trend price he had estimated was $2,208 based on the average of the "'high" year 1987/8, predicted to be $2,536, and the "low" year 1986/7, known to be $1,879. 451. Instead of following the rational course of increasing his trend scrap price, and decreasing the trend rebar price to conform with actuality, he, in fact, increased the trend rebar price to $2400, while leaving his original forecast of $514 for scrap undisturbed. 452. While scoring high marks for audacity, Mr Gillett's credibility took a hard knock from doing that. 453. We do not suggest that Mr Gillett sought actively to mislead the Tribunal. What we do feel is that the longer he was retained as an expert witness, the more he warmed to SFI's case. While his First Report (31/01), was, generally, objective and reasonable, his Second Report (31/02), and his performance in the witness-box, revealed he had become a partisan in SFI's cause. 454. His lapse from that detachment the courts seek in an expert witness is the subject of further comment in our Section III : "REBAR PRICE" 455. All that one can say about $514 as the trend price for scrap is that it was the first number Mr Gillett thought of. 456. SFI would have the Tribunal believe that because he is a steel economist who, rightly for some purposes, has been put forward as an expert, it means that, by virtue of his judgmental powers, he can give figures off the top of his head, which the Tribunal should then accept. 457. Certain types of expert witness, in appropriate situations, do not have to back their opinions with reasons or analysis. For example, there is the "look-sniff" type of expert on whether, say, a cargo of vegetable oil is rancid. 458. Economists are in an entirely different situation from that. They cannot reasonably expect courts to accept their judgments except where sensible and underpinned by reason and analysis. 459. For two reasons, we reject Mr Gillett's attempt to cling to his low forecast price of scrap rather than the higher actual. 460. Firstly, Mr Gillett lacks the gift of clairvoyance. It is no more within his power to give accurate prophecies of scrap or rebar prices than it is for, say, a stock-broker or economist to tell you the price of a share, say, six months hence, or for a commodity-broker to try to do the same for the price of gold. Unless backed up by analysis, such opinions are virtually worthless at the best of times. 461. There was no analysis of any substance from Mr Gillett explaining why the Tribunal should stay with his original forecast of $514 rather than the actual $546. All he gave were vague generalities, and the only particulars he gave related to his knowledge of SWS's prices from July to September 1988, which were in the low $ five hundreds. 462. As we have already indicated, Mr Gillett's first report was made in February 1988 and his second in November 1988 His testimony was also in November 1988. 463. Nothing, in our view, happened between February 1988 and November 1988, which justified Mr Gillett's moving away from his objective method of taking SWS's average scrap price for the year 1987/8. What he said originally in his First Report (31/01, page 38) turned out to be valid in relation to SWS's actual scrap price of $546 for 1987/8 : "1987/88 average is likely to be neither a very high or very low year for scrap prices in Hong Kong". One only needs to glance at SWS's average scrap prices from August 1980 to April 1989 to realize that (See Exh SF 211 at pages 49, 50 and 51). 464. The second reason why we will not allow Mr Gillett to change from his objective method to his subjective opinions was the view we formed that he had lapsed from the comparative detachment of his First Report. 465. What struck us as probably the fairest way of arriving at a trend price for scrap was to take a simple average of prices for the five years 1983/4 to 1987/8. That could be done with either SWS's or KYM's annual average prices. As we have already concluded that KYM represented the market price for cash-collect scrap in Hong Kong, it follows that, for the sake of consistency, it is KYM's average we should take. The result would be $559 (plus 10% stemming from our "two-in-the-market" finding). It was Mr Meocre Li who pointed out the appropriateness of such an approach for a situation where there is no obvious trend. If Mr Li had pressed us to adopt that approach, we would have done so. 466. Instead, he seemed content to settle for a lower figure - $541, which was KYM's actual cost of scrap for Financial Year 1987/8. That, too, embodied a rational approach, namely, treating scrap as a manufacturing cost like any other, such as e.g. electricity or melting materials. Then the actual price of scrap in 1987/8 is used, along with all the other 1987/8 costs, for the Loss of Profit calculations for 1988/9 and beyond where all the figures are to be in 1987/8 dollar values. 467. As Mr Li did not press for a higher base figure than $541, that is the figure we adopt for the scrap trend price, plus, of course, 10% (for the "two-in-the-market" point), which produces the result $595. 468. A point we note here is that there is no necessity for the scrap "trend price", which has to be in 1987/8 dollar values, to be the same as the 1987/8 actual price. For example, the approach we favoured-most, namely, a simple average over five years would have resulted in a figure ($541) in the 1987/8 accounts different from the trend figure ($559 plus 10%) for 1988/9 and beyond. 469. It was not just because Mr Gillett's trend figure of $514 differed from his actual 1987/8 figure that we rejected it. SECTION 2 -SUMMARY OF RESULTS
Notes: (i) Net cash flow from January 20, 1987, in the year 1986-87. (ii) No adjustmenmt made for annual decrease in the depreciation allowances used in calculating the taxation charge & liability in this and subsequent years. (iii) Per John E. Medtey of Mclellan and Partners Limited. ------------------------------------------------------------
Notes (i) A nominal rate of 33% per annum has been used in 1987-88, as inflation has been taken into account in the estimation of profits for these years. (ii) Cash flows are discounted to January 19, 1987, the date the Land was vacated. (iii) It has been assumed that the cash flow accrues evenly during a year. Cash flows have therefore been discounted from the mid point of each period.
*working capital at July 1, 1986 of 11,000, as per SF226, plus Section III, REBAR PRICE 470. It was common ground that for Financial Year 1981/2 - the last year before the price SFI got for its rebars started to suffer the effects of "the shadow" - the income SFI received from the sale of its rebars worked out at $1644 per metric ton. 471. As a fact, we find that the volume of its deliveries to its customers that year was not affected by "the shadow". Beyond mere assertion from Mr Gillett, the steel-market economist called on behalf of SFI, there was no evidence to support SFI's contention that in the No-Scheme-World that year, it would have sold 65,000 metric tons of rebar, rather than the 61,607 M/T it in fact sold in the Scheme-World. On the view we take, the evidence fails to show the threat of resumption having any measurable effect on SFI's 1981/2 sales in the Scheme-World. 472. Thus, SFI's actual volume of-deliveries that year, namely 61,607 metric tons, when multiplied by the delivery price of $1,644, results in income from rebar sales of $101,282,000 for 1981/2. (See DHS's Revised's Restated Profit and Loss.Accounts, at page 023 from their Bundle dated December 1991, being Appendix II, but noting that the description in the margin "Sales - standard length steel bars" is inaccurate, since the figure of $1,644 per MIT includes income from all bars, whether standard length or non-standard.) 473. In Arthur Andersen's equivalent Appendix II in their Bundle of Revised Appendices of 18th December 1991, Alternative 2; Arthur Andersen has correctly shown under "PRICES PER TON", "Sales-inc special, length steel bars". 474. For the sake of completeness, we mention that the figure $1,644 also includes receipts for mild steel, imperial measures, non-standard diameters and other variations (if any) from "standard" rebars, "standard" being high-tensile, deformed, reinforcing bars of twelve metres in length, with diameters ranging from 10-40 mm and conforming with BS4999. 475. For the Financial Years 1982/3 to 1987/8, the Tribunal has to determine "No-Scheme-World" earnings which SFI could reasonably have expected to receive in each of those years. 476. For the purpose of the present exercise, we focus on the delivery price per metric ton that we think SFI would have received year by year for its rebar. All the rebars it'would have sold in this era and onwards are of the high-tensile, deformed, variety meeting BS4999. 477. We find as a fact, for reasons to be mentioned later in this Section, that 75% of its sales each year were of standard sizes, namely, 12 metres long and of diameters in the 10-40 mm range. Those standard sizes fetch the base price, e.g., as we shall show in a moment, our base price for 1982/3 is $1,578 per metric ton. 478. The remaining 25% of its sales giving rise to income each year would have been of non-standard sizes, which is to say, longer or shorter than twelve metres, and/or of diameters outside the 10-40 mm range. 479. Those non-standard sizes earn SFI 6% more than the base price, so we find. Bearing in mind that only 25% of the bars SFI sold attract that additional 6 per-cent, it therefore receives, overall, 1 1/2% above the base price. Thus, to amplify our example, its average earnings per metric ton for 1982/3 work out at $1602 (i.e. $1578 x 1.015). 480. On the basis of our finding of fact that SFI'S production capacity for 1982/3 was 82,000 M/T, we find it would have had gross receipts of $131,364,000 that year. At this point, we are only mentioning this for the purpose of illustration, and, in due course, will show what we found SFI would have earned in each of the years 1982/3 - 87/8, and why. In respect of the "why", we will analyse the alternatives urged upon us by the parties for determining SFI's income for those years, SFI championing a method based on an official government index published monthly, known as the Average Wholesale Price Index for High Tensile Standard-Size Reinforcing Bars whilst, on behalf of the government, it was argued that the actual prices received by Hong Kong's only other mini-mill, SWS, from deliveries arising from its sales of rebar year by year to SFI's sister company, the building contractors, Hip Hing ("HH"), should serve as the measure of what SFI would probably have earned per metric ton for its rebar in a shadowless world. Both of those approaches will be fully ventilated in due course. 481. For the Financial Years 1988/9 and onwards, there is agreement that the income will be expressed in 1987/8 dollar values, the mechanism to achieve this end being a trend price per metric ton of rebar earnings for 1988/9 and onwards denominated in 1987/8 constant dollars. Thus, starting with Financial Year 1988/9, this flat, inflation-free trend price, carries on. All items, whether income or expenditure, in SFI's Profit and Loss Accounts beginning with Financial Year 1988/9, belong to an inflation-free model agreed to by the parties. 482. The rebar trend price we have found is $2,250, for reasons on which we elaborate in due course. As will be shown, that figure includes the 1 1/2% premium for special lengths. That implies a base price of $2,217 per M/T. 483. In view of our finding that SFI's production capacity settled down at 110,000 M/T per annum in 1985/6, and then onwards, SFI's gross income from rebar deliveries in 1987/8 constant dollars for Financial Years 1988/9 onwards will be $247,500,000 ($2,250 x 110,000). 484. We now descend to details. 1982/3 to 1987/8 485. Generally speaking, we think the First Report (31/01) of Mr Gillett, the steel market economist retained on behalf of SFI towards the end of 1987, was a highly commendable effort. 486. On an amorphous mass of facts and figures, he imposed order, and introduced much light. That First Report was both scholarly and lively, as one would expect from someone who had headed the Industry Section of the Economist's Intelligence Unit for seven years, besides having served in several other positions which will have contributed to his expertise as a steel-market economist. We have noted that, in particular, his experience has an international dimension. 487. For that First Report of his, he hit upon the idea of using a yearly A.W.P.I., based on the months of SFI's Financial Year, as an indicator of the prices at which SFI might reasonably have been expected to sell its rebars in the No-Scheme-World. 488. That index has been officially published by the government's Census and Statistics Department ("C. &S.D.") every month since early 1977. It shows what it describes as the "average wholesale price" for reinforcing bars in the month preceding publication. 489. Initially, the same index was divided into a column for mild steel reinforcing bars, and another for high-tensile. Since about 1980, mild steel has been removed from that particular index, and is dealt with elsewhere in the monthly price data the government publishes on different types of building materials. 490. The A.W.P.I. for high-tensile steel rebar shows, for the month preceding publication, what it describes as the "average wholesale price" for "standard length" (i.e. 12 meter bars, in the 10-40 mm diameter range). 491. The information making up that index is culled from thirteen or so - some parts of the evidence suggest as many as twenty but nothing turns on the precise number - wholesalers of rebar, who are approached by officers of C. &.S.D. on the lst, 11th and 21st of each month for the prices at which they sold volumes of ten tons or more of standard length high-tensile rebars on those days. Before resumption, SFI itself was one of the wholesalers which contributed data. 492. The prices collected in this way will include contract and spot sales, and will relate to large as well as small volumes, down to the minimum of 10 tons just mentioned. 493. There was a suggestion that the same sort of data on standard length high-tensile rebars was collected by C. & S.D. from 13 or so contractors to be include in the computation.of the A.W.P.I. Some doubt was cast on that by Mr Widdicombe in his closing address, when he gave the impression that the information from the contractors was used solely for a different government index known as the Lands & Works Index. 494. Certainly, as we understand the situation, the data from the wholesalers, which gives rise to the A.W.P.I., also finds its way into the Lands & Works Index, but whether the contractors' data also find its way into the A.W.P.I. is unclear on the evidence, but, as nothing turns on whether the contractors' data gets fed into the A.W.P.I., we see no reason to make a finding on that, and consign it, with the thousand-and-one other loose-ends in this case, to limbo. 495. There appeared to be no dispute that the reason C. & S.D. collects its data for the A.W.P.I. is to facilitate the operation of escalation clauses in government building contracts 496. For reasons known best to itself, the average calculated by C. & S.D. for the A.W.P.I. is a "simple", not a "weighted" average. The significance of the difference between simple and weighted averages loomed large in the present case. 497. Simple averages work out very nicely for the contractors with the benefit of escalation clauses in their contracts with government, and, no doubt, a great howl would be heard from them if the government, without consultation, were suddenly to switch to weighted averages as a substitute. 498. An illustration, based on the examples (e.g. Exh R88, R90 & R97) given by Mr Meocre Li, a partner in Arthut Andersen & Co., the accountants on behalf of the government, will illustrate the difference between the two types of average. Wholesaler X sells 1000 MIT of rebar at $2,000 per MIT, while wholesaler Y'sells 50 MIT at $3,000 per ton. If a simple average is taken of X and Y's prices, one gets, the answer of $2,500 per ton ($3,000 + $2,000 divided by 2). 499. A weighted average, however, which takes account of the volumes involved in each of the transactions, gives a very different answer :
500. Dividing the total price of $2,150,000 by the total tonnage of 1050, one arrives at the weighted average price of $2048 (corrected to the nearest whole number), per ton. 501. Most helpfully, Mr Li rang the various changes involved in weighting the A.W.P.I., showing how, not only there could be weighting by volume in respect of the 13 wholesalers making returns of data to the C. & S.D., but, also, how, as a preliminary step, each one of those 13 wholesalers could have been the object of a separate weighting exercise, so that if, say, a particular wholesaler on 1st July had sold 100 tons at $2,000, 250 tons at $1,900 and 500 tons at $1,800 the correct average by weighting for volume would,be $1,853, while the simple was $1,900:
502. Thus, if our imaginary wholesaler chose to report to C. & S.D. that he had sold 850 tons at an average price of $1900, the A. P. I. resulting from that data would finish up somewhat skewed for anyone seeking to find out comprehensively what the selling price per ton really had been that month. 503. In practice, there is not likely to be much difference in times of stable rebar prices between an index based on simple averages and one which is weighted. With everyone selling at more or less the same price, there will be little difference to show up. 504. It is when the market overheats, and prices go berserk, that the difference between weighted and simple averages in the context of an index becomes significant. 505. Such a market occurred during Financial Year 1987/8, when the price of high-tensile standard length rebars shot up from $1829 per M/T in August 1987 to $3043 in February 1988, from where it flopped back to $2563 in June 1988. We heard, for example, how, in February of that particular Financial Year, some spot sellers of rebar were getting as much as $4,000 per ton. Because such sales are likely to have been of small volumes, they will have a distorting effect on an index based on simple averages, wholly disproportionate to the volume involved in that transaction. Mr Gillett colourfully described that sort of transaction as "rogue tonnage". 506. It so happens that the Financial Year 1987/8 is particularly important for the purposes of our case, since it immediately precedes the Financial Year when the model agreed by the parties moves over to inflation-free trend prices for scrap and rebar, along with inflation free prices for all other items in the Profit and Loss Accounts. 507. Financial Year 1987/8 is also the last year for which we have some evidence of an actual mini-mill's earnings, namely, the income S.W.S. received for its deliveries to HH. As with earlier years, one can contrast how much per ton HH had to pay SW for its deliveries that year ($1942, see SFI 216, page 154, Co1.4) with the A.W.P.I. price per ton that year ($2458 see Exh.R10A (Revised), Col.3). 508. Another figure of some relevance for comparison purposes is the weighted average price per M/T at which SWS agreed during that same Financial Year to sell standard length high-tensile rebars to HH under long-term contracts. The figure there is $2234, and for that, we have to thank Mr Best, a partner in Deloitte Haskin and Sells, SFI's accountants for the present case. Mr Best caused such weighted averages to be calculated for the Hip Hing/Shui Wing long term contracts from Financial Year 1982/3 when HH and S.W.S. first started doing such business together, uptil and including 1988/9: See Exh SFI 216 at pages 142 to 149, and, more precisely, p.147 for 1987/8. 509. As an aside, we feel it might be helpful at this juncture to point out the importance of distinguishing delivery prices from contract prices. In Mr Gillett's First Report (31/01), for example, where he refers to rebar prices he usually means delivery prices. Thus, when he noted that SFI's prices in 1981/2 were higher than the A.W.P.I., he, in context, meant delivery prices. Elsewhere in evidence, however, he and other witnesses when referring to SFI's prices often mean contract prices. In effect, therefore, "price" sometimes means income already earned, and at other times refers to future income. We mention this in the hope of eliminating that particular pitfall. 510. Having now touched upon the topic of SFI's delivery prices and weighted average contract prices for rebars, we can re-focus our attention on Mr Gillett's First Report 31/01. As we mentioned earlier, he there pioneered the proposal of using a yearly A.W.P.I. as the measure of SFI's income from rebar sales in the No-Scheme-World. 511. We do not doubt for one moment that, when advancing that proposal, he genuinely believed in its appositeness for the task in hand, and put it forward in a constructive spirit as the answer to the conundrum of what SFI would have earned in the No-Scheme-World. 512. As he rightly observed, the A.W.P.I. is a professionally collated, officially published index: No one doubts its reliability. To borrow a label Mr Gillett applied in a different context, the A.W.P.I., on first encounter, beckons as a "dens ex machina". 513. We, too, confess to having been initially attracted by its Siren-song, but now recognize it as a god that failed. 514. The A.W.P.I. has all the virtues claimed for it, except relevance. When used for its correct purpose of being a component of the Lands & Works Index, it is an irreproachably objective yardstick. However, when prayed in aid as an indicator of SFI's revenue, its limitations can quickly be exposed. For example, in Financial Year 1978/9 (See Exh R10A revised), the A.W.P.I. showed $1585 per M/T; SFI's actual earnings that year were $1214 per M/T. 1979/80 shows a similar disparity, the A.W.P.I. producing a figure of $1893, while SFI's actual takings were no more than $1588 per M/T. 1980/1, the gap closes with the A.W.P.I. $1728, and SFI earning $1704. 81/2 there is a reversal with the A.W.P.I. at $1624 trailing SFI's receipts of $1644. Then the "shadow" falls. 515. A vast collateral exercise was mounted on behalf of SFI in an endeavour to explain away the lack of congruence between the income SFI had actually received per metric ton in 1978/9 and 1979/80, and the A.W.P.I. for those two years. It was urged upon us that the answer lay in a constellation of circumstances such as, (i) a greater volume of orders being placed with SFI in the first part of calender years 1979 and 1980 when the A.W.P.I. was relatively low compared with the second half of those calendar years; (ii) the lag between placing orders under long term contracts and actually receivigg the income; (iii) the difficulties of pricing in a rapidly rising market; (iv) the possibility of SFI's earnings, at least in 1978/9 and 1979/80, and perhaps even later, being depressed by receipts from cheaper mild steel getting mixed in with SFI's takings for high-tensile during those years; (v) Imperial sizes in those early years possibly having had a similarly depressing effect, and (vi) there was the circumstance that SFI did not start to get any revenue enhancement from special lengths until at least 1979, the first year SFI entered into a contract making provision for a premium for special lengths. 516. On each of those six points we will make comments in due course. 517. Suffice it to say at this point that, if. Mr Gillett's hypothesis of the A.W.P.I. measuring SFI's income needs shoring up by first of all trying to explain away the miscellany of circumstances we have just referred to, it loses the merit of simplicity which initially made it so beguiling. 518. Based on what we regard now as the simplistic notion that the annual A.W.P.I., synchronous with a Financial Year under the "shadow", was a reasonable measure for SFI's loss of revenue from rebar sales, for that Financial Year, Mr Gillett, in his First Report, (31/01), went on to say that SFI's financial performance could be expected to be even better than a Financial Year's A.W.P.I., because of the uplift SFI would get for its special lengths. He conservatively estimated the premium for special lengths as half of one per cent, a figure at which the government did not cavil. 519. Ultimately, SFI had to make concessions, in the face of overwhelming evidence, that an undiluted application of the A.W.P.I. would overstate SFI's loss of revenue. For the Financial Years 1978/9 to 81/2 inclusive; (the latter year being the last before the "shadow" began to take effect), SFI had contracted on the basis of a simple average price for those years at 1.28% below the A.W.P.I. (We think a weighted average should have been used resulting in the figure of 1.82% below the A.W.P.I. See Exh SF216 at p.71 and Exh R155B). 520. On behalf of SFI, the suggestion was in due course made, and we accept its logic, that, as a step towards reconciling contract and delivery prices, a contortion has to be performed lagging the A.W.P.I. by four months, the agreed average mid-point of SFI's long-term contracts, after allowing for cash sales. 521. Yet a further contortion is called for, in our view, if consistency is to be achieved on how SFI matched its sales contracts with the A.W.P.I. SFI contended that unnecessary distortion arose unless a contract in a particular. month was compared with the A.W.P.I. for that month. (See SFI216 at page 71 A-C) For the government, the less refined approach of comparing a whole year's A.W.P.I. with a whole year's contracts was advocated. 522. On that particular feat of arithmetical gymnastics, we thought SFI clearly carried off the medal. 523. Abandoning Mr Gillett's uncontroversially conservative figure of one half of one per cent for special lengths, SFI, in the end, said it wanted 1.66% (based on approximately 27% of its sales being special lengths which commanded on average a premium of 6.22%). As previously indicated, we have found as a fact that the figure for the special lenghts' premium is to be a rounded 1 1/2%. 524. A far cry from the attractive simplicities of Mr Gillett's First Report is SFI's final position on calculating the loss by staggering the A.W.P.I. for four months, discounting that by their historic 1.28% average price below the A.W.P.I. and then adding back on 1.66% for special lengths. 525. SFI,still relies on those convolutions which can be found worked out in, e.g., Mr Best's set of accounts dated August 31, 1989 in his volume 33/06 at page 47, but the Tribunal has, also, been presented with a no frills alternative, dated November 15, 1989, at page 47A, which lacks the lag, and the 1.28% taken off followed by the 1.66% put back on. (See C.C.S. section 4, Vol I, Part I, blue pages 19 and 20, and para 4 of SF270.) 526. What we chose, and why, will be described in due course. 527. Three prongs of the government's attack on the A.W.P.I. - based approach to revenue, recommended for SFI by Mr Gillett, can conveniently be considered together, in view of the significant ways in which they overlap or interact. The three issues raised can be summarised as (a) whether contractors generally time the placement of their rebar contracts in a way which anticipates significant upswings in the A.W.P.I. or whether their timing is random; (b) whether SFI's income in relation to the A.W.P.I. should be weighted over a period of years, (this being a different question from whether the A.W.P.I. itself should be weighted) and (c) the effect of the lag between SFI's delivery-based income and the contract-based A.W.P.I. Each of these three issues involves a different facet of the effect of timing on SFI's earnings. 528. Potentially, the most devastating of all Mr Meocre Li's forays against SFI's contention that the A.W.P.I. stood as a fair measure of its loss was his theory that contractors were able to read the market in such a way as to enter into long-term, high volume, contracts with wholesalers like SFI a short time before the market made a mjaor upswing. By long term contracts Mr Li, in context, was referring to contracts of the order of six months to two years. 529. As empirical support, Mr Li could point to how contracts for 43,748 M/T's (See SF 216, page 166) had been placed with SFI between January and June 1978 when the A.W.P.I. kept within the range $1,021 to $1,280 per MIT (See SFI 216, page 160). During the following six months, the A.W.P.I. steadily mounted to $1,466, and contractors now stayed their hand, ordering a mere 6,314 M/T's, whereupon Mr Li's point (b), about weighting over a period of years, gets brought into play. Although the A.W.P.I. was moving up nicely from July to December 1978, i.e. the first part of Financial Year 1978/9, SFI is seen not to be benefitting from the rise to the extent it clearly would have hoped, since it has only managed to get contracts for a comparatively small volume in the good times. The vast flood of relatively low priced orders in the earlier January to June 1978 period exerts a downward pressure on SFI's income from deliveries right into Financial Year 1979/80, by when the A.W.P.I. is in the $1,800 to $1,900 per M/T range. 530. It will be self-evident how the weight of low priced deliveries in Financial Year 1978/9, and even beyond, stemming from the contracts made from January to June 1978, will have opened up a sizeable gap between SFI's delivery price in 1978/9 - $1,214 - and the 1978/9 A.W.P.I. of $1,585 (I have not overlooked mild steel, Imperial measures etc., etc., but understood SFI was not pretending that particular gap could be explained away entirely by those sorts of exotica). 531. As the weighting point is important, we feel it might be as well to re-produce the example (Exh R96) Mr Meocre Li helpfully worked out for the Tribunal, as it is comparatively short :-
532. Mr Meocre Li also produced a rather clever exhibit, (R98), to further illustrate his theory on how the timely placement of large, long-term contracts for rebar could enable contractors to get deliveries at prices far below the A.W.P.I. at the time of delivery. All along Mr Li made it completely clear that he had introduced some simplifying assumptions into R98, and it was only an illustration, not a purported description of actuality. 533. Further empirical grist to Mr Li's mill was furnished by SFI's contracts made in January, February and March 1979 to sell 21,237 M/T (See SFI216, page 126), a period during which the A.W.P.I. stood between $1,466 and $1,738 per M/T. In the half year from July to December 1979, a period during which the A.W.P.I.'s high-point reached $1,912 and low-point $1,843, only 8,777 M/T in contracts came SFI's way, so, again, there was the phenomenon of half a year with a high volume of sales s at comparatively low prices for the A.W.P.I., followed by six months of low volume and a high A.W.P.I., leading to another chasm between SFI's delivery prices for 79/80 of $1,588 and the A.W.P.I. for the same period of $1,893. 534. Needless to say, as always, the inevitable, routine time-lag between long term contracts and the actual deliveries under them - Mr Li's point (c) - was at work in . 1978/9 and 1979/80. As there is, generally, - at least in nominal terms - an upward trend in rebar prices, that lag tends to put the A.W.P.I. above current delivery prices. On rebar price trends, we will have occasion to say more in due course. 535. Notwithstanding the considerable spread for the Financial Year 1978'/9 between the $1,214 per M/T SFI earned for its deliveries that year and the A.W.P.I. Of $1,585, the weighted average price per ton, (calculated SFI's way in accordance with the approach described in SFI 216 at pages 71A to C), at which it contracted to sell its rebar that year was a mere 0.14% (SF1216, page 71) below the A.W.P.I. The following Financial Year (1979/80), when.SFI received $1,588 per M/T for its deliveries, the A.W.P.I. worked out at $1,893 per M/T. That year SFI contracted at 0.32% below the A.W.P.I. Despite thus contracting within a hair's breadth of the A.W.P.I. for those Financial Year's 1978/9 and 1979/80, SFI's earnings for each of those years, expressed as a percentage of the A.W.P.I., worked out at 76.2% and.83.5%, respectively. 536. That is demonstrated in Exh R105 After 1979/80, the correlation between SFI's earnings and the A.W.P.I. for the same Financial Year improves substantially, reaching 98.1% for 1980/1, 100.7% for 1981/2 and then, under the "shadow", 97.2% for 1982/3. 537. Not even the agreed average four month time-lag between contract and delivery does all that much to help SFI, at least in the earlier years. With that lag built into the calculation done by Mr Li in Exh R89, one sees that for Financial Year 1978/9, SFI's delivery prices turn out to be 87.8% of the four-month - lagged A.W.P.I., compared with 76.2% without any lag. For 1979/80, the comparable figures are 84.5% lagged, 83.5% unlagged. For 1980/1, it is 94.4% and 98.1%, on this occasion SFI making a worse showing lagged than unlagged. 538. For 1981/2 and 1982/3, the lagging bring SFI'S delivery prices to within almost 100% of the A.W.P.I., the precise figures being 1981/2: 99.7%, and 1982/3: 99.7%. 539. In the face of Mr Li's clear demonstration of how, in Financial Years 1978/9, 1979/80 and 1980/81, the A.W.P.I. was a hopeless indicator of SFI's earnings for those years, even when the A.W.P.I. was modified by a four month lag, one finds the following assertion in CCS, Section 4, vol. I, Part I, blue page 17, para. 5.2: -
540. That is simplistic, and seeks to perpetuate Mr Gillett's fallacy on how the A.W.P.I. is a good measure of SFI's income. Paragraph 5.2 is only correct when confined to contracts made in a single year, say, for example, 199X. If one follows.up the deliveries from those contracts, whether taking place in the contract year 199X, or in the year or years (i.e. 199Y and 1992) subsequent to the contract year, yes, the revenue from those deliveries will amount to the contract price multiplied by the volume ordered. However, in the real world, there is the ongoing process of deliveries from contracts in different years (e.g. 199V, 199W and 199X) being received in the same year (say, 199X), so it is unlikely that the A.W.P.I., whether lagged or-uniagged, will be a good indicator of income for any particular year, except in times of stable prices, or except where, unlike SFI in 1978/9, 1979/80 and 1980/1, the seller not only manages to keep his contract prices abreast of the A.W.P.I., but also avoids the trap of selling relatively large volumes under long term contracts when the A.W.P.I. is low, but relatively small volumes when high. 541. A variant of the same fallacy can be found in SFI 216, page 109, para. 11.8:-
More appropriate opening words would be "In theory", rather than "In practice". 542. From Exh R105, one sees that for the four Financial Years preceding the effects of the "shadow" (i.e. 1978/9 to 1981/2), SFI's delivery prices were, on average, a little over 10% short of the unlagged A.W.P.I. for those years, while Exh R89 discloses that with a four month lagged A.W.P.I., the gap narrows to, approximately, 7 1/2%. 543. Allowing for the whole string of reasons (e.g. mild steel, Imperial lengths et al.), advanced on behalf of SFI, concerning why its delivery price performance relative to the A.W.P.I. for that four year period was-not-a good harbinger for what SFI could reasonably have expected to earn in the No-Scheme-World, Mr Li voiced the opinion that the delivery price would, all the same, during that period 1978/9 to 1981/2, have been about 6% short of the A.W.P.I. 544. He went on to assert that,' whatever the reasons for that historic 6% shortfall, the A.W.P.I. per metric ton for the time under the "shadow" should be trimmed, for reasons of empiricism, by that same 6% to arrive at a delivery price per metric ton which could be treated as a fair measure of income foregone by SFI during the "shadow". 545. Before reaching the point where we move on to consider the validity of Mr Li's theory of major rebar contracts being timed in such a way as to be placed, at or near the foot of major upswings in the A.W.P.I., we will first of all attempt to assess the significance of SFI's efforts on the six grounds already mentioned to explain why its earnings remained so far short of the A.W.P.I. in the years 1978/9, 1979/80 and, (when lagged), 1980/1, and why there should be no fear of history repeating itself in the No-Scheme-World. 546. Firstly, it was SFI itself which drew attention to the circumstance that, in the first six months of each of the calendar years 1978 and 1979, long term contracts for high volumes had been placed at times when the A.W.P.I. was low compared to the remaining six months of each of those calendar years. In a way, SFI picked up a rock to drop on its own foot by pointing that out, since Mr Li not only agreed with.it, but seized upon it to fashion his theory of contractors placing long-term, high volume contracts at the foot of major upswings in the A.W.P.I. 547. For reasons we will give after the present exercise of considering SFI's six grounds for why pre-shadow history need not be repeated post-shadow, we do not accept Mr Li's hypothesis that, as a general rule, contractors can be expected to place high-volume, long term contracts in time to benefit from upswings in the A.W.P.I. As we will in due course explain more fully, we regard the fact that contractors did place those large contracts in the first halves of calendar years 1978 and 1979 as owing more to happenstance and luck than any superior predictive skills on the part of the contractors. We are not persuaded that there is some iron law of economics which condemns a wholesaler like SFI over the years to receive rebar income perpetually falling about six percent below the A.W.P.I. 548. On the other hand, we do not consider that SFI would have acquired immunity from the workings of chance in the No-Scheme-World, and do not think the circumstance that HH placed long term contracts for higher volumes with SWS in Financial Year 1986/7, when the A.W.P.I. was $1879, than in the immediately following Financial Year 1987/8, when the A.W.P.I. was considerably higher at $2458, as a valid reason to forswear the adoption of HH/SWS delivery prices as a reasonable measure for SFI's likely loss of income in the No-Scheme-world. We regard it as inherently preferable to follow a real-life surrogate for the simulation of SFI's likely income, rather than a much-massaged A.W.P.I. as requested by SFI, adding on 1.66% here, taking off 1.28% there, and so forth. We have no illusions that the SWS/HH analogue is perfect, and we will have more to say later about its shortcomings, but, in our view, it is far and away the best surrogate we have. Had it survived shadow-free in the real world post 1981/2, SFI would have had good years and bad years for earnings relative to the A.W.P.I., like any other wholesaler, and we see no reason to suppose that, overall, (apart from special lengths), its fortunes could be expected to have been significantly different from those of SWS, as revealed to us in its dealings with HH. 549. During the Financial Year 1981/2, the tonnage SFI contracted to sell for the six months July to December was 34,576 M/T, while the amount for January to June of 41,262 was not disproportionate when compared with the difference between SFI's volumes of contracts in the first halves and second halves of earlier years. The position can be seen from the following table lifted from SFI's Exh 216 at page 166. Table 1 SFI Total Sales Contracts
550. From that, we see the great disparity from one half year's volume to another's, until we come to Financial Year 1981/2. 551. According to the submission made on behalf of SFI, a new era had dawned in 1981/2, and from then on, had it not been for the "shadow", there would have been little imbalance between contract volumes from one half year to the next in the No-Scheme-World. That state of affairs was supposed to flow from the circumstance that with SFI's increased production capacity, there would usually be contracts in every month and a more even flow of tonnage. 552. While we accept that, in the No-Scheme-World, as in the Scheme-World 1981/2, SFI would probably have had contracts in every month, that does not give rise to any inference that the tonnage contracted for in one half year will be of more or less the same order as the previous half year and the succeeding half year. 553. In calendar year 1978, SFI.had contracts every month but, all the same, the volume of the first half of that year of 43,748 M/T dwarfed the second half's 6,314 M/T. In Financial Year 1978/9, SFI had contracts in eleven of the twelve months, but the difference between the first half of that year 6,314 M/T and the second 27,312 M/T as conspicuously large. 554. There is no reason why, in the No-Scheme-World, there should not be a glut of orders in one half year, followed by a relative famine in the next. We regard it as unreasonable to suppose that, if SFI were offered attractive long term contracts with a volume of, say, 80,000 M/T in the first half of a year in the No-Scheme-World, it would turn down,. say, 30000 M/T. of that, in order to be able be enter into contracts for more or less the same volume in the second half of the year. In reality, we think, SFI would have juggled with its orders from one half year to the next, making sure it could always sell up to its full capacity, but not caring whether there was symmetry in the contracts from one half year to the next. 555. The more or less balancing volumes in the first and second halves of 1981/2-appear to be the result of chance factors, in our view, and do not provide an adequate foundation for SFI's theory about equal volumes of contracts per half year in the No-Scheme-World. 556. In any event, unequal volumes of contracts from one half year to another are of no particular significance in times of stable prices, such as occurred throughout the years 1979/8 to 1986/7,except for the period May to October 1983. 557. The second of SFI's six grounds for contesting Mr Li's assertion that, in the light of its pre-"shadow" history, SFI in the post-"shadow" era should receive no more than 94% of the A.W.P.I. per M/T for its deliveries, centres on the importance of appropriate lagging of the A.W.P.I. for the purpose of gaining an understanding of its true relationship to deliveries. 558. In an endeavour to illustrate how, amongst wholsalers of rebars in Hong Kong, not only SFI's, but also SWS's and Tung Wing Steel's (TWS), delivery prices over. the years on average fell several percent below the A.W.P.I., Mr Li produced Exh R105, being one of a series of what he described as ".reality tests". We produce a copy of the table of figures making up that exhibit: "COMPARISON OF STEEL BAR DELIVERY PRICES
NOTES: 1. The SF actual prices include special length bars 2. Tung Wing actual prices are calculated from the Tung Wing prospectus (RL-68) 3. HH/SW actual prices and average wholesale prices have been adjusted for the special length premium (5% x 10%)" 559. From that, one sees how the figures for SFI's "Actual" i.e. delivery prices, trailed far behind the A.W.P.I. in the years 1978/9 and 1979/80, being years when the A.W.P.I. was rising rapidly. 560. Exh R105 purports to show something similar happening to the delivery prices of both SWS and TWS in 1983/4, another year when the A.W.P.I. rose sharply. 561. Figures are lacking for TWS for the Financial Year 1987/8, which was the next occasion for a rapid rise in the A.W.P.I., but, for SWS that year, we see 78.1% recorded for the delivery price of $1929 per M/T expressed as a percentage of the A.W.P.I. of $2470 per M/T. 562. In Exh R89, an earlier "reality test", Mr Li had, to a considerable extent on that occasion, appeared to cut the ground from under SFI's feet by showing that, at least for 1978/9 and 1979/80, there was still a substantial difference between SFI's delivery prices and the A.W.P.I., even after applying the suggestion made on behalf of SFI that the A.W.P.I. should be lagged by four months to make allowance for the agreed average interval between the making of SFI's long term contracts and making the actual deliveries under them. The figures set down in R89 are as follows: "Shun Fung Ironworks Ltd
Objective of Table : To test the validity of the propositions and conclusions set out in pages 40 to 48 of SF216. Notes: (1) The average wholesale price ("AWP") adopted is the base index and does not include any allowance for special length. (2) The SFI weighted average price includes the premium for special length as the actual basis." 563. In relation to Exh R105, Mr Li did not follow through with lagged calculations on SWS's and TWS's delivery prices. If, however, that exercise is done, the whole picture on the relationship of delivery prices to the A.W.P.I. is transformed in SF-I's favour. 564. We agree.with the reasonableness of the uncontested submission made on SFI's behalf that the average lag between contract and delivery for TWS should be treated as four months and,. for SW/HH, five months. We now show the figures for TWS and SWS/HH, respectively, incorporating those lags, with the same sort of information as in Exh R89. The material for TWS, which is taken from CCS, Sect.4, Part I, Vol 2, Appendix 4 to page 133, Table II and III, is as follows: "Comparison Between AWP and T.W.S. Delivery Prices
We have employed the base A.W.P.I. (i.e. without any addition for special lengths) on this TWS exercise. 565. As those TWS delivery prices relate to calendar years, the four-month-lagged A.W.P.I. corresponding with calendar year 1987 terminates in August 1987, which means it just misses, the meteoric rise which-began with the September 1987 A.W.P.I. and continued through until February 1988. Therefore, there is no opportunity to see how TWS performed during that critical period (hence the gap in Exh R105), but Table III is still useful for the time of the A.W.P.I.'s more restrained advance in 1983/4. For SWS/HH, we have adapted SFI's material from the same volume of the C.C.S., Appendix 3 to page 133, Table I, by basing the A.W.P.I. only on the months in which SWS and HH contracted together. That is consistent with the approach explained by SFI in SF216 at pages 71A to C, of which we have already expressed approval. As with the TWS table, we show the A.W.P.I. as a base index, i.e. without any addition for special lengths. To the SWS/HH delivery prices for standard lengths. (SF216, page 154, Column 2), we have added 1 1/2% to represent the premium which we have found, as a fact, SFI would have earned for special lengths in the No-Scheme-World. Our table is as follows:
566. An alternative way of presenting the same information, so as to follow the form of Table II (at page 8A) of Appendix 3 to page 133 of CCS, Sect.4, Part I, Vo1.2, is as follows :
567. On the basis of the latter table, we have calculated that for the period 1982/3 to 1987/8 inclusive, SWS's average receipts from deliveries to HH, assuming a 1 1/2% premium for special lengths added to the price for standard lengths, would have worked out at 98.71% of a five-month-lagged A.W.P.I. made up of the months in which SWS/HH had long term contracts (i.e. the contract months shown in Exh SFI 216 from pages 142 to 147). 568. As we have already noted, the Financial Year 1987/8 was exceptional for the way in which the A.W.P.I. increased by approximately 66% between August 1987 ($1,829) and February 1988 ($3,043). 569. If one excludes that exceptional year when calculating the average - The ground we will be coming to next is the difficulty of pricing in a rapidly rising market - one finds that for the period 1982/3 to 1986/7, the average becomes 99.95%. 570. The same calculations can also be done on the basis of a four month lag, the agreed average interval between SFI making long term contracts and effecting delivery of the rebar. By some quirk, that works out less favourably to SFI than with the SWS/HH five month lag, at least for the full 1982/3 to 1987/8 period, for which the answer is 97.92%. For the 1982/3 to 1986/7 period, SFI would gain slightly on a four month lag approach, the result being 99.98% compared with the previously mentioned 99.95% for five months. 571. For the four month lag, we now set out a table following the form of Exh R89, and the alternative in the form of Table II (at page 8A) of Appendix 3 to page 133 of C.C.S., Sect.4, Part I, Vol 2:-
and
572. The full workings of our calculations on both the five-month (Table 1) and four-month-lag (Table 2) basis are shown at the end of the present Section. 573. Much play was made, on SFI's behalf, of the fact that SWS had entered into contracts with HH in only 29 of the 72 months of the period from July 1982 to June 1988. 574. It was pointed out for SFI that, during that time, 2808 returns would have been made to the C. & S.D. for the purpose of compiling the A.W.P.I. 575. According to the argument advanced for SFI, the results of 2808 returns must be better than 29. This brings us back to the now-familiar terrain of weighting, and the sector of the market where SFI would have done business in the No-Scheme-World. Based on those considerations, we are of the view that the 29 HH/SWS long term contracts are a more suitable indicator of SFI's income in the No-Scheme-World than the A.W.P.I. made up from the unweighted returns of the thirteen wholesalers, most of whom would be too small to operate in the part of the market where large volumes are sold under long term contracts. That part of the market is the preserve of the likes of SWS, TWS, and, formerly, SFI. 576. Because of the stability of the rebar market throughout most of the period up to August 1987, it turns out that, until then, it hardly makes any practical difference whether one uses SFI's recommended approach of the A.W.P.I. suitably lagged, or the SWS/HH delivery prices for standard lengths plus 1 1/2% for the extra SFI could have expected to earn for special lengths. 577. As we have shown, for the years 1982/3 to 1986/7, the A.W.P.I. with the five month lag appropriate for the SWS/HH transactions at 99.95% comes to within a whisker of the SWS/HH delivery prices plus the 1 1/2%.supplement for special lengths. With a five-month lagging of the A.W.P.I. uptil August 1987 - the last month before the A.W.P.I. commenced steep climb - the SWS/HH delivery prices should remain a good approximation of what SFI might have expected to earn from its deliveries in the No-Scheme-World over the five months September 1987 to January 1988. We regard SFI’s complaints about the use of the SWS/HH delivery prices before then as of academic interest only. Their points about the smallness of the sample, the uniqueness of the revenue of any one wholesaler depending on the price, timing, and length of particular contracts, and the SWS/HH contracts not being representative of SFI’s market appear to have been of no practical consequence from an earnings point of view until February 1988. 578. It is because we have acceded to one of the government's submissions that we have selected the SWS/HH series of transactions as the basis for measuring SFI's loss of income from rebar sales. Having persuaded us to follow SWS/HH, the Crown has to, we feel, accept the full consequences of that, including the circumstance that the lag appropriate to those transactions should be five months We recoil from the prospect of the sort of hybrid which might result from crossing the SWS/HH transactions with the SFI delivery period. The principal attraction to us of using the dealings between SWS and HH as a surrogate is their origin in a real-life situation, - a reason that makes us reluctant to allow any tinkering. The approach we have endeavoured to adopt rests on impartially requiring each side to accept the consequences, good and/or bad, of the SWS/HH transactions serving as surrogate, with no liberty to blow hot and cold, and without endless massaging of the figures backwards and forwards. 579. A few paragraphs back, we foreshadowed the third of the grounds put forward by SFI to explain the gap between the A.W.P.I. and its delivery prices pre-"shadow", namely, the difficulty of pricing in a rapidly rising market. 580. Mr Roy Leung mentioned that problem in paragraph 15.6 of his First Affidavit, made long before he could have known that the government would contend that, in Financial Year 1987/8, that sort of history had repeated itself. 581. It was, no doubt, the events occurring in Financial Years 1977/8 and 1978/9 which prompted Mr Roy Leung to share his views,on the problems faced by sellers of rebar in rapidly rising markets. 582. Although we.can see from SFI 216 at page 160 that in 1977/8 the A.W.P.I. rose from $982 in July to $1,279 the following June, we do not know whether, for that year, Mr Roy Leung overcame the difficulties of which he has spoken, or whether he took a hammering from them. 583. Our only hard data on the contracts SFI made that year consists of the bare, unprocessed information to be found in Annexure M of 42/01. 584. According to Mr Li, SFI was, on average, contracting at 5 or 6% below the A.W.P.I. in 1977/78, but the picture is so lacking in detail, and befogged by the wild-card of cheaper mild steel being mixed in with the dearer high-tensile variety, that we really do not know what happened. 585. In respect of that year, we merely note that there was a particularly sharp rise from January to February ($1,021 to $1,139), and a fairly sharp one from February to March ($1,139 to $1,212). Apart from that, the A.W.P.I. mounted steadily throughout the year. 586. 1978/9 was another year of the A.W.P.I. rising continuously, with sharp spikes from January ($1,466) to February ($1,580), February to March ($1,738), and March to April ($1,884). SFI did well to achieve a weighted average price for its contracts that year 0.14% above the A.W.P.I. for the year. See SF 216 at pages 71, and pages 125 to 127. Constantly rising prices in those two Financial Years 1977/8 and 1978/9 must have helped condition SFI to cope with the difficulties of which Mr Roy Leung spoke. 587. 1979/80 ushered in an era of relatively stable prices for the A.W.P.I. on a month by month basis, with relatively gradual moves up or down. 588. Again, in 1979/80, SFI beat the A.W.P.I., by 0.32%. 589. For the next two years, SFI's weighted average contract prices were below the A.W.P.I. - by 3.03% in 1980/1, and 2.55% in 1981/2, the last year before the "shadow". In the remaining years till leaving Junk Bay, SFI did worse than the A.W.P.I., but that is of little, if any, importance, on account of the effect of the "shadow". 590. The long run of stable prices, as reflected by the monthly A.W.P.I., came to an end in May 1983. April 1983's A.W.P.I. had been $1,620, but then may jumped to $1,718. June was $1,867. The mini-boom straddled the new Financial Year 1983/4, with the A.W.P.I. for July at $1,906, and culminating at $2,192 in October 1983. 591. After that, the A.W.P.I. went through another long, stable phase until Financial Year 1987/8, when, in September, it moved up to $1,889, having been $1,829 in August. Between August 1987 and January 1988 ($3,043), the A.W.P.I. exploded upwards by the 66% of which we have, earlier in this judgment, had cause to note. One would, surely, suppose that would qualify as the sort of rising market in which Mr Roy Leung had said it was difficult for a mini-mill to price its rebars. 592. Looking at how the surrogate SWS/HH performed in the years following the "shadow", we see that the results were not that far different from SFI with all its customers, in SFI's two last shadowless years. 593. As we have said, 1980/1 SFI trailed the A.W.P.I. for the year by 3.03%, and then in 1981/2 by 2.55%. 594. Then, switching to SWS/HH for the succeeding years, we see percentages, all under the A.W.P.I., of a similar order
595. During those particular years while SWS was selling to HH, the only time SWS would have been put to the test by a rapid rise in prices was between April 1983 ($1,620) to October 1983 ($2,192). 596. For May 1983, we see that SWS's weighted average contract sales price managed to beat the A.W.P.I. by 2.84% (See SFI 216 at page 142). The only other month in which it contracted during that period was July, when its price, on the same basis, fell 3.71% below the A.W.P.I. 597. Out of interest, we can take a look to see how SFI, under the "shadow", was faring during that same period within 1983, and we see it had contracts only in August and September when its weighted average prices were 4.98 and 4.60, respectively, below the A.W.P.I. for those months: (See SFI 216 at pages 76 and 77). 598. Turning now to look in more detail at the SWS/HH contract performance in terms of the A.W.P.I. during the time from August 1987 uptil February 1988 when, as we have already shown, the A.W.P.I. sky-rocketed from $1,828 to $3,043, it is instructive, first of all, to study how the A.W.P.I. had behaved in the months leading up to that. From March to August 1987, we see the A.W.P.I. keeping within the narrow band $1,828 to $1,836. That, in turn, had been preceded by an unbroken period of stable prices stretching back to November 1983. 599. Of the period 1987/6, Mr Mui, HH's manager, said, with what we regard as some understatement, prices for rebar were "abnormal". 600. During the critical period August 1987 to February 1988, SWS made contracts with HH in five of the seven months. 601. In September and October, we see (from Exh SFI 216 at page 147) that SWS beat the A.W.P.I. for those months by 3.24% and 2.57%, respectively. For the other three months, however, - November, January and February - SWS was beaten by the A.W.P.I. to the tune of 7.21%, 14.21% and 11.57%, respectively. 602. For the whole Financial Year 1987/8, SWS/HH contract prices were 5.06% below the A.W.P.I. 603. In C.C.S. Sect.4, Part 1, Vol.l, page 63 paragraphs 12.6 and 12.7, those acting for SFI have introduced another of their wild-cards, this time taking the form that, whereas the A.W.P.I. was based on rebars in the 10 - 40mm range, calculations done on behalf of the government in Exh R40, relating to the SWS/HH contracts in Financial Year 1987/8, have been limited to bars in the 16 - 40mm range, and Table XIII at page 141 of SFI 216, which purports to compare the weighted average price per ton of standard length H.T. rebar in SWS/HH contracts, sometimes could not make full provision for 10 - 12mm rebars, as the volumes were not always stated in the contracts. 604. As 10 - 12mm bars are sold at a premium to the 16 - 40mm bars, the effect of their omission from the calculations means an understatement of the SWS/HH weighted average price per ton for the purpose of comparison with the A.W.P.I. 605. Where the contracts do not show the volume, the effect is, by definition, unquantifiable, and it is, frankly, pointless to be told about it. 606. In the earliest such contract (No 3543 of May 1983), the difference between the 10 - 12mm bars and the larger bars is only $30 per M/T, but in the latest contract (No 4629 of November 1987) was $100 per M/T. 607. As the smaller bars are more expensive to make, since they need more rolling, it is anyone's guess what the effect of these unquantified bars implies for SFI's profitability. 608. Some comfort is also to be found in the knowledge that, even if the higher price of the smaller size bars fails to get reflected in the weighted average contract price per ton, there will be no corresponding problem in the SWS/HH delivery prices which embrace the full 10 - 40mm range. 609. It is pleasing to be able to note, too, that the uncertainty stemming from Exh R40 is, in fact, not so great as those advising SFI fear. Exh R40 can be correlated with Exh SF216 at page 147. Except in relation to contract numbers 4,580, 4,609 and 4,629, the information necessary to calculate the effect of the small bars is there. 610. Mention need only be made of the four largest contracts in terms of tonnage, namely No.'s 4,539, 4,608, 4,618 and 4,689. The following table shows the percentage difference between the SWS/HH contracts and the A.W.P.I., according,to Exh R40 and SFI 216 at page 147:
611. We take the view that such information as is still lacking, after comparing Exh R40 with SF 216 at page 147, should be treated as neutral in its effect, with no practical implications for the case. 612. Nothing in SFI's history, nor its attributes brought to our attention, leads us to regard, it as reasonable to suppose that it would have performed any better than SWS during that extraordinary year when, in Mr Roy Leung's words, there were, a fortiori, "... the difficulties in pricing when selling into a rapidly rising market ..." 613. We now move on to the fourth of SFI's six reasons for the gap between its rebar earnings and the A.W.P.I. for 1978/9 to 1980/1, this fourth reason being that until about 1981/2, part of SFI's deliveries under contracts might have been of mild steel bars which can be seen from the appropriate A.W.P.I. to have sold at a discount to the high-tensile variety. The differences can be seen from the following table extracted from SFI 216 at page 167.
614. On behalf of SFI, it was conceded that the problem need not trouble the Tribunal after about 1981/2, due to the combined effect of the difference in price between mild and high-tensile narrowing to insignificance (in fact, in a few months, the mild, according to the relevant A.W.P.I., was dearer than the high-tensile), and because so little mild bar was sold latterly, in any event. Despite that concession made on behalf of SFI, an attempt was made to resurrect the spectre of mild steel on one occasion in relation to an SWS/HH transaction which, inevitably, had to be from the 1982/3 era onwards, but we ignored it as being nothing more than a red herring. 615. Where any of SFI's contracts made prior to lst January 1980, did not specify the quantities of high-tensile and mild rebar, the Tribunal was, in effect, invited by SFI to assume the ratio of high-tensile to mild bar would be 15:25 uptil 1st January 1980, and 85:15 thereafter. 616. On behalf of the government, no fuss was made about SFI's assertion that those were the ratios until SFI started playing what Mr Carnwath, for the government, rightly described as the "wildcard" of how the gap between SFI's earnings and the A.W.P.I. might be, to some extent, explained away by the possibility of lower prices from mild steel bars being mixed in with SFI's earnings from high-tensile. No attempt was made by SFI to quantify the effect of this mild steel on its earnings for any particular year. For example, we know that SFI's actual delivery price per ton was $1,214 in 1978/9. With the possibility of cheaper mild steel hovering in the background, what are we supposed to think about that $1,214? How much, if anything, are we supposed to add on to make-allowance for more notional high-tensile rebar to offset the depressing effect of the unmeasured volume of mild bars? We are left perplexed, with the problem dangling in the air. 617. As SFI's quantity ratios of 75:25 pre-lst January 1980 and 85:15 after, rest on nothing more than assertion, and, as for the government, it has been stated those quantity ratios are not accepted as correct, we do not think that SFI has sufficiently laid the groundwork for the nebulous addition it wants us somehow to make to its earnings with a view to narrowing the gap with the A.W.P.I. Uncertainty about the significance of mild bars can cut both ways. There is no proof that the quantity ratio, high-tensile to mild, was 75:25. For all we know, on the evidence as it stands, it might have been, say, 95:5, which would mean SFI had overstated, rather than understated the high-tensile element. 618. We have the feeling of having been sent off on a wild goose chase by SFI over the possibility of mild bars leading to understatement of SFI's earnings. 619. The fifth (rather like the,fourth) of SFI's reasons for its poor earnings relative to the A.W.P.I. in the pre-shadow years was, again, in the nature of a hunt for a chimera. This time, the topic was whether SFI's earnings before the "shadow" might have been depressed from lower priced Imperial sizes of rebar being counted together with higher priced metric sizes. The difference in price between the two types can be gleaned from the following table extracted from SFI 216 at page 168: Average Price Quotations
620. Whether, in practice, Imperial sizes had any effect on SFI's earnings, and, if so, how that effect should be quantified, we simply do not know from the evidence presented to us. 621. Perhaps, as Mr Li said, mild steel and Imperial sizes might have led to SFI's 1978/9 and 1979/80 earnings being 1 to 2% lighter, and we are prepared to accept what he says, despite the inconclusive nature of the evidence on these topics. In any event, we do not think it makes any difference to the outcome of the case. 622. Sixthly, and lastly, there is SFI's contention that its earnings prior to about 1980 have to be seen in the light of no loading at that time for special lengths, the first contract which made provision for special lengths not having been entered into until February 1979, and, even for many months after that, a mere scattering of its contracts made such provision. 623. As previously indicated, we have accepted as a fact that, in the No-Scheme-World, special lengths would have brought SFI a premium of 1 1/2% beyond what it would have earned for standard lengths. Later in this judgment, we will elaborate on our reasons for that. 624. Having finished our digression on the six particular grounds put forward on behalf of SFI to explain why its earning per metric ton delivered in some of the pre-shadow Financial Years fell so far short of the A.W.P.I. for those same Financial Years, we can pick up the threads once more of Mr Li's theory on major rebar contracts being timed in such a way as to be placed at or near the foot of major upswings in the A.W.P.I. 625. The data on which Mr Li based that theory can be seen diagrammatically in Exh R99. 626. It is instructive next to look at the diagram on R97, which is a composite chart, showing, firstly, the volume of SFI's long-term contracts on a month by month basis from January 1978 until the "shadow" took effect towards the end of Financial Year 1981/2, and, then, secondly, the SWS/HH long-term contracts from Financial Year 1982/3 to December 1988, that period 1982/3 onwards, of course, being under the "shadow". 627. There is no suggestion that Mr Li worked out his theory on the basis of any general principle. Instead, his theory is of the a posteriori variety. By a process of inductive reasoning, based on the high volumes of contracts placed in the first half of calendar year 1978 and then in the first half of calendar year 1979, being followed soon afterwards by a rapid rise in the A.W.P.I. on each occasion, he has gone on to propound his theory. 628. The phenomenon of the placing large volumes of contracts being shortly afterwards followed by major upswings in the A.W.P.I. happened on only the two occasions just mentioned, namely, the first half of calendar year 1978, and then the same time the next year. 629. If one confines one's attention solely to the chart R99, which shows only the position of SFI from January 1978 till the effect of the "shadow" towards the end of SFI's Financial Year 1981/2, Mr Li's theory might, perhaps, appear to have some glimmerings of validity, but, even then, there is the awkward circumstance of SFI receiving massive contracts totalling 15,421 M/T in September 81, followed until June 1982 by a gentle descent of the A.W.P.I. 630. In April 1982, SFI got an even larger volume of contracts, 25,213 M/T, being its most ever in a single month. Exh R99 is neutral as to whether the behaviour of the A.W.P.I. after those substantial April 1982 contracts supports or refutes Mr Li's theory : the chart stops in June 1982. When, however, the progress of the A.W.P.I. after June 1982 can be followed on R97, the composite chart showing the A.W.P.I. in relation to SFI contract volumes pre-shadow, and the SWS/HH volumes post shadow, it becomes crystal-clear that Mr Li's theory simply will not hold water. 631. After April 1982, the A.W.P.I. is seen not to start rising again until a year later. 632. In November 1983 (14,015 M/T), April 1984 (12,000 M/T), and January 1988 (18,000 M/T), SWS received contracts for those large amounts at times more or less coinciding with peaks in the A.W.P.I. 633. In February 1985, HH signed contracts with SWS for 21,181 M/T, but the A.W.P.I. does not start to ascend until August 1987. 634. We formed a very high opinion of Mr Li as an expert witness. Regarding most of the other manifold issues on which he expressed opinions, we thought his evidence was characterised by rationality and restraint, but, in relation to his theory of large contract volumes being followed by steep climbs in the A.W.P.I., he had a blind-spot, as he is demonstrably wrong. 635. Before turning to consider the government model, based on the HH/SWS delivery prices for rebars as indicators of SFI's income in the No-Scheme-World, we are, first of all, going to consider the significance of both SFI's and SWS's contract prices for standard lengths frequently turning out to be less than the A.W.P.I. (which is, of course, also based on standard lengths) for the month in which the contract was made. That shortfall was generally described throughout the case as a discount to the AWP, but it is important to bear in mind that, at the time the contract was made, no one knew how the A.W.P.I. was going to turn out for that particular month. Neither SFI nor SWS would say to a customer "We will give you a discount of X% to this month's A.W.P.I.". 636. It is only when C. & S.D. get around to publishing the A.W.P.I. in the following month that the parties to the contract get the opportunity, with hindsight, to see, if they are interested, whether the contract price turned out to be at a discount or a premium to the A.W.P.I. 637. From both Mr Mui of Hip Hing and Mr Lam of E Man, as well as from the evidence generally, it was clear that neither contractors nor wholesalers resort to the A.W.P.I. when negotiating a price for rebars. 638. Mr Gillett, and the submissions made on behalf of SFI, kept repeating that the A.W.P.I. was the Hong Kong market price for rebar. On such evidence as we have, it was certainly not the price for the sector of the market where SFI and SWS, Hong Kong's only mini-mills, dealt. It was empirically demonstrated that, on average, over the years, they were contracting at close to 2% or more below the A.W.P.I., the precise figure for SFI in respect of all customers for the 1978/9 to 1981/2 period on a weighted, average basis being 1.82% below (see Exh R155B), and for SWS/HH, on a simple average basis for 1982/3 to 1988/9, being 2.62% below (see SF216 at page 141). 639. Both of them turn out, more often than not, from the evidence before us, to have contracted below the A.W.P.I., when one compares a contract month with its A.W.P.I. 640. Even in times of relative price stability such as, say, 1980/1 and 1981/2, we see how SFI's weighted average contract prices for standard lengths fell below the A.W.P.I.- by 3.03% and 2.55%, respectively. The same happened with SWS/HH contracts in, say, the stable years 1985/6 and 1986/7, when their weighted average contract prices were 2.99% and 3.18%, respectively, below the A.W.P.I. 641. Then, as Mr Widdicombe very pertinently asked, who are the people getting the high prices which push the A.W.P.I. above SFI's and SWS's contract prices? 642. Could it be TWS No, we do not think so. TWS's delivery prices, on such evidence as we have, stand in the same relationship to the A.W.P.I., appropriately lagged, as those of SFI and SWS with their respective lagging, so it is not unreasonable to presume that TWS's contract prices would have started out with "discounts" to the A.W.P.I. not so different from those of the other big players, SWS and SFI. (We have not overlooked the point that TWS did not come upon the scene until 1983/4, by when SFI was in eclipse). 643. SWS and TWS between them captured a huge share of HK's rebar market. Mr Roy Leung said, at one point, SWS had 70%. We think that figure for SWS is somewhat on the high side, and believe the figure bandied about by others of about 45% was probably nearer the mark. 644. TWS's share we accept to have been 35% of the market, so that, between them, SWS and TWS had about 80% of the market. 645. As just indicated, even in stable times, SWS usually contracted below the A.W.P.I., and, judging from such indirect evidence as we have, based on delivery prices, TWS's contract prices were unlikely to have done better. 646. By a process of elimination, it looks as if it probably was some of the 11 or so smaller wholesalers who between them shared the remaining 20% of the HK rebar market. Stemming from the quirks of simple averages and the way in which the A.W.P.I. is calculated, some relatively high priced transactions, involving small tonnages, whether on a spot or long-term basis, could distort the A.W.P.I. in a way which made it unrepresentative of the experience of either SFI or SWS. 647. During the period from August 1987 to February 1988 when the A.W.P.I. went haywire, rising from $1,828 to $3,043, we see from SWS's contracts with HH during that time that they were as far as 14.21% below the A.W.P.I. (for January), but never more than 3.24% above (for September). 648. From Mr Gillett and others, we heard of prices as high as $4000 per M/T being paid on the spot market that February. The price, however, we see SWS agreeing with HH that month is only $2691 (the A.W.P.I. turning out to be $3043). 649. Some of those smaller wholesalers must have taken advantage of the run-away market at that time to do some price-gouging, but, as far as we know, SWS, judged from its dealings with HH, did not. We doubt very much if it did with other customers, either. 650. From all we have heard about SWS, we formed the impression it probably shared the rather likeable "dinosaur" characteristics, claimed by Mr Roy Leung for SFI, of reacting somewhat slowly to events. 651. We tend to doubt that a huge, long established company like SWS would have played the spot market to extract every last dollar. Instead, SWS would probably have gone on fostering long term relations, treating its customers fairly and decently, in the same way SFI did. 652. We regard it as reasonable to suppose SWS, even during that boom period, charged all customers the same. That was the way SFI had behaved, as was pointed out to us on their behalf. When there was a shortage of rebar on the market, the problem for the small contractor was getting a contract at all from the big wholesalers, but, if you were large contractor of the type which was in the market for high volume contracts, the large wholesalers would still enter into contracts with you at reasonable prices because of your muscle. 653. As Mr Mui put it, the large contractors had the upper hand in dealing with the large wholesalers. SFI’s sector of the rebar market was the one where the large contractors, wanting to assure the availability to themselves of the high volumes of rebar needed over long periods of time for huge projects, dealt with the large wholesalers who could meet their requirements. The attraction to a mini-mill of entering into those sorts of contracts was that they allowed the mini-mill to plan long production runs. Mr Roy Leung made that point time and time again. 654. One sees that in 1981/2, the last year before the "shadow" took effect, roughly 90% of SFI's long term contracts were with just three large contractors : Hip Hing, E. Man, and Sin Chong.. Theirs were obviously the contracts which determined the level of SFI'S prices which, as a matter of SFI's policy, were the same for everyone. SFI had many customers besides those three, but as everyone was asked to pay the same, the circumstance of a large number of customers is of no significance in the pricing context. We accept, too, that about 20% of all SFI's sales were on a spot cash basis, but, as their price, was the same as for the contracts, they have no impact on SFI's overall prices. 655. In April 1982, so we were shown, SFI had its best ever month for volume, selling 25,213 M/T to five different customers, the largest sale being for 13,305 M/T and the smallest 484 M/T. Each of those five customers paid the same, namely, $1,571 per M/T, which turned out to be 2.10% below the A.W.P.I. for that month. 656. On behalf of SFI, the submission before us proceeded on the assumption that, each and every customer, in the No-Scheme-World, would at all times have been paying top dollar, but the reality we see, when reviewing how SFI had in fact behaved, is that its policy of uniform prices for all customers, in practice, meant a levelling down rather than up. 657. Not all wholesalers could have been as ethical as SFI, in dealing with contractors. Hence, the data finding its way into the A.W.P.I. which brings out the point that SFI was not getting the highest prices. Because SFI's prices did not vary with volume does not mean the position was the same with the smaller wholesalers. 658. It is during a time like the overheated market between August 1987 to February 1968 that the significance of the weighting of averages on a price/volume basis comes into its own. One of the eleven small wholesalers contributing data to the C. & S.D. who managed to sell, say, 100 metric tons for, say, $3,500 per MIT in January 1988 carries the same weight in the A.W.P.I. for that month as SWS which, we know, sold HH 18,000 MIT that month for $2,537 per MIT. A simple average of those two figures works out at $3,019 per MIT, but, weighted, the average becomes $2,542.
659. Whether one chooses to attribute the difference between the two numbers to weighting, or to the effect of a discount, does not really matter. What counts, for present purposes, is the inability of an index such as the A.W.P.I., calculated on the basis of simple averages, to give a fair indication of revenue for a large wholesaler such as SFI. 660. Although we do not believe that anyone has the ability to predict rebar prices, that does not, however, mean that people in the business are unable to recognize the frothing of a wildly overheated market when it arises. 661. Between August 1987 to February 1988, there developed a shortage of rebars in Hong Kong. If you were a small contractor, carrying no clout with the big wholesalers (i.e. SWS and TWS), but, desperately in need of rebars, you were vulnerable to being squeezed for an outrageously high price by the small wholesalers. In that situation, you are small, and the wholesaler is small, but your transaction (which is also small) influences the A.W.P.I. for the month in question to the same extent as a massive, long-term contract between SWS and HH. When SWS and HH negotiate that contract, SWS does not say to HH, "We are going to give you a discount of 15% below the A.W.P.I. due to be published next month" What happens is that SWS and HH agree on a price together which seems reasonable in the light of all they know about the market in Hong Kong including the circumstance that, given a little time, imports will flood back into the place, thus restoring the mark'et's equilibrium. Part of the background knowledge against which SWS and HH will negotiate their contract might be that some unfortunate small contractor, had the other day, paid $3,500 per M/T in the spot market, but that does not mean it would be irrational for SWS and HH to agree on a price of, say, $2600 per M/T for their long term contract, nor would it even be irrational for SWS to agree at the same time to sell some bars to HH for immediate delivery at $2600 per M/T, bearing in mind their long standing relationship. In acting that way, SWS's behaviour would be rather similar to what Mr Roy Leung told us SFI used to do. 662. In the extreme circumstances of the overheated market between August 1987 and February 1988, it is easy to detect the limitations of simple averages. 663. Although not so obvious at times when the market is calmer, the potential for simple averages to lead to unfairness is still there. As already mentioned, the simple averages of the A.W.P.I. would have led to unfairness in 1980/1 and 1981/2 if the A.W.P.I. for those years had been taken as an indicator of SFI's prices under all its contracts which were, once weighted, in fact 3.03% and 2.55%, respectively, below the A.W.P.I. 664. We cannot help feeling that, by putting so much emphasis on the A.W.P.I., SFI has lost sight of the fact that, the A.W.P.I. in the present case, is only a means to the end of ascertaining SFI's loss, and is not an end in itself. 665. That confusion about means and ends was at its most obvious when Mr Gillett described how he had set about trying to predict the A.W.P.I. for 1988/9 and beyond, whereas the question to which he should have been directing his mind was what SFI's earnings for the future were likely to have been. 666. Another point which needs to be said about the A.W.P.I. is that, although it is compiled from market data, it is, nevertheless, fallacious to go on from there to claim that the A.W.P.I. must therefore be the market price. The problem is semantic. In some contexts, "market" means the undifferentiated market as a whole: that is what a reasonable man would have in mind when reference is made to the "market data" collected for the A.W.P.I. "Market data" in that sense embraces the spot market, the short end of the market, the long end of the market, mini-mills, large wholesalers and small wholesalers. 667. When referring to SFI's "market", the context obviously means the sector where SFI does business. 668. The issue which next falls to be considered is whether the government has been right in asserting that the SWS/HH delivery prices are the fairest measure for SFI's loss of earnings in the No-Scheme-World. 669. There can be no doubt that Hong Kong contractors prefer to buy their rebars from a local mini-mill rather. than from a wholesaler-stockist. An unsolicited testimonial to that effect is found in HH's answers made to a questionnaire from the Hong Kong Productivity Council in 1980 at a time when there was not the slightest hint that resumption lay in store for SFI. (Document Code No 21/26) 670. At the time of answering that questionnaire, the only mini-mill of which HH had experience was its sister-company, SFI. 671. SWS was already in a substantial way of business then, but had not yet started its own melt-shop. It had rolling mills for which the feedstock was imported billets, and it also imported rebars for sale. 672. SWS's meltshop commenced operations in 1980, but it did not start to rank HH among its customers until 1982 when, as a matter of policy prompted by the "shadow", HH, at the instigation of its parent, NWD, started switching its orders away from SFI to SWS. 673. Until then, far and away the greater portion of HH's rebars were ordered from SFI, but when the prospect arose of SFI's Junk Bay factory being forced to close, it made no sense for the sister companies, SFI and HH, to risk perishing together. 674. We accept that all along it had been NWD's policy that HH had to pay SFI the full market price for any rebars it bought. For reasons we have already explained, that did not mean HH had to pay SFI the equivalent of the A.W.P.I., which is an ex post facto, simple, unweighted average of the prices of wholesalers big and small, on transactions big and small (down to 10 M/T). 675. The sector of the market SFI targeted was the one concerned with large volume, long-term contracts, and it was the price appropriate to that market which HH, and SFI's other customers, had to pay to SFI. 676. Partly because of their special relationship as sister companies, and, partly also in recognition of how advantageous it was to HH to get its supplies from a local mini-mill as flexible as SFI, which, as a fact, we find, had built up a niche for itself in special lengths, HH was frequently willing to pay SFI slightly over the odds. Instances where this might happen, so we accept from Mr Mui's testimony, were where HH had called for tenders from rebar suppliers : in awarding the contract to SFI, HH might agree to pay SFI slightly more than the lowest of the bids. 677. However, HH was in no position to be too open-handed, since it was operating in a highly competive environment, and had to cost its materials very carefully. 678. If we adopt SWS/HH delivery prices as the surrogate for SFi's lost earnings, the problem of making allowance for SFI's additional income from its special niche of special lengths presents no difficulty : one merely adds on an appropriate percentage, which, as a fact, we have found to be 1 1/2%, to the price of the SWS/HH standard length bars. 679. But how about adding on something further in recognition of the little extra Mr Mui told us HH sometimes gave SFI? 680. There appears to be only the one year where HH can be seen showing conspicuous generosity towards SFI: that was 1979/80 when HH contracted to pay SFI prices for standard lengths which, on a weighted average basis for that year, were 4.68% above the A.W.P.I. 681. However, the impact of that 4.68% gets diluted if, as we think should be done, SFI's prices are weighted over the four year period from 1978/9 to 1981/2, after which the effects of the "shadow" get felt. That weighting results in an average of 1.12% below the A.W.P.I. for that pre-shadow period, compared with 0.03$ above on a simple average. To us, it seems plain commonsense that a weighted average presents a more informative picture than a simple average. That (1.12%) was arrived at by applying the methodology of Exh R155. We now set out in tabular form the information from which we have derived the (1.12%) "Comparison of Weighted Average Price Per Ton of SFI/HH Standard Length HT Rebar Sales Contracts with AWP
682. In the No-Scheme-World, SFI would have found itself competing with SWS which had the commercial advantage of being able to ring the changes between importing bars and/or billets if local scrap was too dear, or melting when local scrap was attractively priced. This meant SWS could be very competitive in its pricing. 683. We dismiss, as wishful thinking, that SFI at some foreseeable point in the No-Scheme-World would act as a stockist or import billets, when cheap, for rolling. True, SFI had extended its bar-store at Junk Bay so that, physically, there was no reason it should not act as a stockist. The problem for SFI lay in its parlous financial situation. It lacked the capital to become a stockist, and, bearing in mind how NW allowed SFI to hobble along in the period prior to the "shadow", we see no reason to conclude that, more likely than net, capital would have been made available to SFI for this purpose. It would have had to have been from its own cash-flow that SFI funded activities as a rebar stockist. 684. Merely because SFI had extended its bar store to beyond what it needed for its own maximum production of 110,0.00 M/T per annum from its melting does not lead to the inference it would soon be acting as a stockist. 685. One of the defects in the management of SFI was an inability to coordinate its expansion programme as illustrated by, e.g., not getting a high volume of production from the E.A.F.'s installed by 1975 until the concast, installed in 1978, was made to work properly in 1981. 686. The reality is, too, that if SFI had the ability to trade as a stockist, it could have done so in the Scheme-World, renting premises for the purpose once SFI was forced out of Junk Bay. 687. In the Scheme-World, with SFI more or less limping off the field from mid 1982 onwards, SWS became the only mini-mill operating in Hong Kong. 688. HH, as we have seen, liked to deal with a local mill. 689. As SFI from about mid-1982 generally ceased taking on contracts of greater duration than about six months, SWS became Hong Kong's only mini-mill willing to take on contracts exceeding six months' duration. Naturally, that put SWS in a stronger position, commercially, than it would have been in the No-Scheme-World, where there would have been the constraint of not getting too far out of line from SFI's prices on long-term contracts, and, vice-versa, SFI would have been affected in the same way. 690. In the No-Scheme-World, neither mini-mill would have experienced any difficulty selling all it could produce for a market as ravenous for rebar as we know Hong Kong's to have been, but, in the Scheme-World, SWS must have been able to drive a slightly harder bargain once SFI no longer counted as a rival. That "slightly harder" is not easy to quantify. 691. In the same way, it is difficult to quantify the little extra HH was prepared, sometimes, to pay SFI in recognition of the special relationship in which the sister companies stood to each other, so, likewise, is it difficult to gauge the edge in pricing SWS must have enjoyed when it became, to all practical intents and purposes, Hong Kong's sole mini-mill. 692. On a "swings and round-abouts" basis, we consider the SWS/HH delivery prices reflecting SWS's price advantage in the Scheme-World, when it became Hong Kong's only fully-functioning mini-mill from 1982/3 onwards, will, more likely than not, adequately compensate SFI for the slightly better price SFI could, sometimes, have expected from HH, compared with SWS's prices, at a time when both of them would have been competing in the No-Scheme-world. 693. For SFI, it has been somewhat sweepingly asserted that SWS was so anxious to win HH as a customer that it was willing to offer contracts at prices which were variously described as "below market price", "below the price at which SFI would have accepted", or "below the A.W.P.I.". 694. On such evidence as we have relating to how HH and SWS came to do business together, we do know that it was, initially, HH which set out to woo SWS. HH decided to do that some time in the first half of calendar year 1982 when it became known that, due to the threat of resumption, the future of SFI as a reliable source of rebars, was in jeopardy. No doubt, SWS was highly receptive to those overtures at the time, as the business relationship between SWS and HH would be mutually advantageous. SWS must have shared the public knowledge of the threat hanging over SFI, and must have known why HH wanted to establish business relations. 695. Not only as a matter of common sense would it appear that SWS would have been under no pressure to tempt HH with prices below normal for the corner of the rebar market where the large wholesalers and the big contractors did business together, but one can see from the contract prices themselves, agreed to by SWS and HH, that they were of the same order as those observable between SFI and HH in the pre-shadow years. 696. Consider the first contracts SWS and HH made together in October 1982. The weighted average contract price for the 9,586 M/T involved was $1,597 per M/T, which, for what it is worth, turned out to be 0.07% above the A.W.P.I. when, in due course, it was published for that month. 697. Those October contracts were made at a time of relative stability in the A.W.P.I., the figures per M/T for the three months preceding October being $1,596, $1,611 and $1,610, and for the three succeeding months $1,558, $1,546 and $1,548. 698. For the whole of Financial Year 1982/3 when SWS contracted to sell HH 12,941 M/T, SWS's weighted average price per ton of standard length rebar in those contracts turned out to be 1.06% under the A.W.P.I. for the year, a percentage which compares favourably with SFI's weighted average contract prices with HH in the two last years before the "shadow", 1980/1 being 2.91% under the A.W.P.I. that year, and 1981/2; 1.70% under. 699. The 12,941 M/T SWS sold HH in 1982 represented 33% of HH's rebar requirements that year. 700. For the following Financial Year, 1983/4, SWS contracted to sell HH 38,261 M/T, being 59% of HH's requirements that year. Mr Gillett, at the foot of page 5 of his Second Report, (31/02), has this to say about what transpired between SWS and HH that year
701. SWS's weighted average price per ton that year in its contracts with HH had turned out to be 2.94% under the . A.W.P.I. 702. What Mr Gillett is asking the Tribunal to infer from that is that SWS had engaged in price cutting to win over HH's custom, with the result that the SWS/HH price would not be a fair proxy for the prices SFI could reasonably have hoped to get in the No-Scheme-World. 703. Because SFI, in fact, contracted with HH in 1980/1 at 2.91% below the A.W.P.I., it would not be sensible to start suggesting SFI must have embarked on a policy of price-cutting that year, and it is no more rational to make a suggestion along those lines about SWS in 1983/4 on the basis of numbers thrown up by the A.W.P.I. for that year. 704. On the view we take, the average level of the SWS/HH contracts relative to the A.W.P.I. over the period 1982/3 to 1987/8, in the Scheme-World, fairly reflects the level SFI could reasonably have expected to reach with all its customers during that same period, in the No-Scheme-World. The ensuing delivery prices also fairly reflect the earnings SFI could reasonably have expected in the No-Scheme-World during that period. 705. On behalf of SFI, an attempt was made to show that it was unfair to use the SWS/HH delivery prices as the measure of SFI's loss, since HH paid SWS less for deliveries than it did to other sellers. We have followed the evidence and submissions on this through the meandering trail to be found in Exhibit "MKC3" of Mr Mui's first affidavit, Exhibit SF 216 at pages 116 and 154 (in particular column 4), and CCS Section 4, VO1.1, Part 2, pages 72 to 74. 706. No clear inference arises. One year the suppliers to HH, other than SWS, charge more than SWS, another, less. Further confusion is introduced by SFI raising the possibility that mild steel might affect the result, although our understanding is that, by the time SWS started doing business with HH, mild steel was no longer of any practical significance. 707. On such evidence as there was, the table at the top of page 74 of CCS, Sect.4, Vol.l, part 2 should be modified as follows:
708. The column on the left is from "MKC3", and, on the right, is from SFI 216 at page 154, including the amount SWS received from HH for actual special lengths. 709. In furtherance of the submission that it was unfair to SFI to treat SWS's delivery prices as a satisfactory proxy, an attempt was made to show that during the period from Financial Year 1982/3, when SWS first started to have contracts with HH, until 1985/6, when SFI's mini-mill ceased to operate, SFI's prices for deliveries to HH were generally higher than those of SWS to HH. A perusal of SFI 216 at pages 150, and 154, together with Exh R95, left us with the impression the evidence was so ambiguous, and so many special factors were at work, that no conclusions could be drawn. In 1982/3 SFI's delivery price to HH was 1.14% higher than SWS's, but in 1984/5 it was SWS which was 1.64% higher. In 1983/4 and 1985/6, when SFI'S delivery prices exceeded those of SWS by 10.95% and 15.70%, respectively, special factors were so obviously at work that those freak results have no bearing on how SFI might have been expected to perform in the No-Scheme-World. 710. There was also one month under the "shadow" - December 1982 - when SFI and SWS each had a contract with HH. That month, the A.W.P.I. turned out to be $1,546. SFI's contract in respect of 3,598 M/T was priced at $1,517 per M/T ie. 1.59% above the A.W.P.I. SWS, by contrast, only managed $1,500 per M/T on its contract for 1,700 M/T. That was 2.99% below the A.W.P.I. 711. From that isolated instance, the Tribunal was invited to infer that SFI could sell above the market price, but SWS only below. 712. In our view, that isolated instance can lead to no general conclusion about the respective selling capabilities of SFI and SWS, or about the willingness of HH to be generous towards SFI. There had been, and there were to be again, occasions when SFI's selling price to HH was more than 2.99% below the A.W.P.I. for the month in question, and there were also to be times when SWS's contract price with HH exceeded the A.W.P.I. by more than 1.59%. Perhaps we should also just repeat that we do not regard the A.W.P.I. as synonomous with the market price for either SFI or SWS. FINANCIAL YEAR 1988/9 AND ONWARDS -TREND PRICE 713. Part of the model agreed to by the parties was that the price for rebars and scrap in the accounts for 1988/89 and onwards should be in 1987/88 dollar values. That price for rebar or scrap was referred to throughout the trial as the "trend price". 714. On our view, a trend price is simply a reasonable price for rebar to be used in the accounts for 1988/89 and onwards. 715. Mr Gillett, in his First Report, (31-01), made in February 1988, came up with a current trend value of $2,208 for the Hong Kong rebar price. That was in the part of his Report where he evaluated SFI’s project of establishing a mini-mill in Shunde County, China. For the purpose of that exercise, he assumed that once the mill in China started production, its sale price would be the same as the Hong Kong wholesale price (less freight from Shunde to Hong Kong where the rebar was to be sold). The figure of $2,208.00 we have just mentioned was to be the price of 1 metric ton of standard length high tensile rebar. 716. Mr Gillett calculated his current trend value of $2,208.00 by taking the average of'what he described as the "high" A.W.P.I. year, 1987/88, which he estimated would be $2,536.00, and the "low" year, 1986/87, which was $1,879.00. 717. In that First Report of his, Mr Gillett made it abundantly clear that the steel industry, both in Hong Kong and worldwide, had been through a very lean period, starting from about 1975 and going through until 1986/87. 718. From his First Report, there emerged a picture of the Hong Kong economy generally having experienced a difficult period from about 1982 up to 1985, but, after that, there had been a good recovery here, particularly in 1987. 719. On the topic of rebars, his Report explained that the overall prospects for the building and construction industry in Hong Kong were sound. The implication was, that augured well for rebar manufacturers here. 720. At the time of that Report in February 1988, the most recent A.W.P.I. montly figure Mr Gillett had was for October 1987, when the A.W.P.I. stood at $2,085.00. The most recent information in that Report was to the effect that there had been a shortage of rebar in Hong Kong in early 1988, and that had sent the price shooting up by about 40% to $3,400.00 on the spot market. The cause of this surge of rebar prices in Hong Kong was, he had explained, strong demand here, coupled with a reduction in imports to Hong Kong, as there had been revived demand overseas, particularly in the United States and Japan. 721. The tone of that First Report from Mr Gillett was very positive on the prospects for rebar manufacturers in Hong Kong, and elsewhere, at the time of his report. Hence, in the No-Scheme-World, the prospects for SFI were very good. 722. As events turned out, Mr Gillett's estimate for the A.W.P.I. for 1987/88 was wrong. By itself, there is nothing significant in that. Predicting rebar prices is extremely difficult. Everyone, including Mr Gillett himself, agreed on that. As he put it, "A forecast is just a forecast". Whereas in February 1988, he had forecast the A.W.P.I. for Financial Year 1987/8 would be $2,536.00, it turned out to be $2,458.00, i.e. $78.00 less than he had predicted. That translates into $8.58 million per annum for a mini-mill with 110,000 tons capacity. It turned out, also, that he was wrong in his prediction for the Hong Kong scrap price, based on the S.W.S. price, for that Financial Year : his forecast had been $514.00, but the actual price was $546.00. That $32 difference, in turn, represents almost 4 million per annum ($3,978,304) expenditure for a 110,000 M/T-capacity mini-mill, which requires 124,322 M/T's scrap per annum. 723. In the light of what the 1987/8 actual results of the A.W.P.I. and the S.W.S. scrap price turned out to be, one, at first blush, might have expected Mr Gillett to lower his trend price for rebar and increase the trend price for scrap. Instead, he did the reverse, raising his trend price for rebar to $2,400.00 and sticking to his original prediction of $514.00 for the scrap trend price. 724. Had Mr Gillett continued with his proposal that the trend price for rebar should be the average of the "high" year 1987/8 and the "low" year 1986/7, the result, in the light of the actual A.W.P.I. for 1987/8, would have been a rebar trend price of $2,168.00 (i.e. $1,879.00 plus $2,458.00, divided by two). 725. What made Mr Gillett forsake his original method of taking the average of a "high" and "low" year, in favour of an approach which smacked of plucking the figure of $2,400.00 out of the air? 726. Mr Gillett's Second Report (31/02), which reached the conclusion that the trend price for rebar should be $2,400.00, started off with an "Introduction" to the effect that Mr Gillett had obtained "new information", not previously available, which led him to increase the trend figure from $2,208. Again, on page 5, he referred to "new information". A close reading of 31/02 reveals, however, that there is scant, if any, new information. 727. Moreover, this Second Report goes on to mention, at least twice, that the part of the First Report evaluating the project for SFI to reopen in China was "preliminary". However, there is nothing in the part on the PRC, or anywhere else, in the First Report (31/01), which indicates that the Report was "preliminary". 728. That suggestion from Mr Gillett that his First Report was "preliminary", was part of an attempt to lay the ground work for increasing the trend price of rebar. In his testimony, Mr Gillett sought to explain that the part of his First Report which dealt with the PRC was written in a hurry, and was not well thought out, as he had not realised at the time that the current trend value of $2,208.00 he gave then would carry so much weight for the purpose of ascertaining SFI's loss of profit for a period which might be as long as 10 years, that being the time it might take for SFI to reopen and start making profits in China. 729. We regard the reasons he gave in his testimony for recanting on his original rebar trend price of $2,208.00 as tending towards the feeble. 730. At page 7 of his Second Report (31/02), after referring to his "preliminary market study" and "preliminary forecast" of the Hong Kong price of rebar for 1988/89 onwards, Mr.Gillett refers to "The course of events this autumn ..." having contributed to his decision to raise his rebar forecast. 731. No matter how hard one searches in that Second Report, there is precious little to find by way of events that autumn which could have justified Mr Gillett's decision to increase the rebar trend figure. 732. When giving his evidence on why he no longer stood by his original trend price of $2,208.00, Mr Gillett contended that he had not done enough "analysis" then. At least, in his First Report, there was analysis to the extent that he identified the "low" year of 1986/7 and the "high" year of 1987/8, and expressed the opinion that the average of those two figures should be taken as the current trend value. For the new figure of $2,400.00, the only analysis which Mr Gillett did, as far as we can make out, was regression analysis in order to find a trend line based on the A.W.P.I. from June 1971 to June 1988. 733. Unfortunately for Mr Gillett, his attempt at regression analysis (in exhibits SF 34 and SF 40) turned out to be something of a fiasco. On Exh SF 34(C), he misunderstood the scale, which was at two years intervals. He mistakenly believed a trend line reading of approximately $2,354, which he had wrongfully attributed to August 1988, meant that the trend price for the full Financial Year 1988/89 was already well on its way to $2,400. A correct reading of the chart reveals that the trend price for the full Financial Year 1988/9 was that figure of $2,354.00. (See Exh R110, para.3) 734. We accept that Mr Gillett formed his opinion that the trend price for rebars to be used in the accounts for 1988/9 and onwards was $2,400.00 before he had recourse to regression analysis. His regression analysis was intended to corroborate his trend price of $2,400.00. 735. His hypothesis of a trend price of $2,400.00 was not, however, supported by the data he put forward in exhibits SF34 and SF40. According to our understanding of scientific method, if a hypothesis is not validated by the data, the hypothesis is discarded. On the SFI approach, however, the data get discarded. 736. After Mr Gillett's attempt at regression analysis looked to have all the makings of an own goal, the Tribunal was given to understand by Mr Best and from the claimant's closing submissions that regression analysis was a more or less worthless method which was no substitute for the opinion of an expert such as Mr Gillett. 737. Although this might be somewhat inconvenient for SFI, there can be no doubt at all that when Mr Gillett gave his evidence, he put forward regression analysis as a source of powerful support for his thesis that $2,400.00 was the correct trend price, and described it as a fairly standard thing to do when looking for a trend in economic data. He referred (Mr Gillett's Transcript, page 414, line 16 and page 415 line 1) to how he was relying on this more "rigorous" data in his trend line analysis to support his contention that rebar prices showed a strong upward trend. 738. Despite all Mr Gillet's references to "new information" in his Second Report (31/02), there appeared to be nothing new, apart from the ten most recent months' data from the A.W.P.I. which he had incorporated into his regression analysis (and, maybe, the rocketing rebar price in the EEC, on which we will say more later). 739. In the context of seeking to justify his trend price of $2,400.00, Mr Gillett remarked (his Transcript, p.296, line 19), "Anyone can do that through that sort of data, confirms me in my belief that there is an upward trend in the prices in Hong Kong as in elsewhere, and if you look at that line (meaning the trend line in Exh SFI 34(c), our parenthesis) I don't think you can do anything else but arrive at the conclusion that 2208, my original forecast, is really too low, looking at the 1988/89 onwards, and if you want to put a better figure on it, I would put 2400." Mr Gillett later observed (at page 300, line 3 of his Transcript) "I am saying if you look at my graph (meaning SFI 34(c), our parenthesis) you are looking at the 1988/89 trend figure". A remark along the same lines is made by Mr Gillett a little further along in the transcript (at page 304 line 36) "Could I just say that for me to be consistent, I repeat would be for me to take the 1988/89 trend figure if you like off this graph (meaning SFI 34(c), our parenthesis) if I decide that this is the most representative and reliable way of doing it, and I would take the 1988/89 figure and continue that forward, and that would be consistent with what I did before." 740. Those three last answers of his touch upon another issue of considerable importance in the context of trend prices. When Mr Gillett purported to read off figures for trend prices from his trend line in exhibit SF34(C), he was obviously reading off figures in nominal dollars for whatever the year was, and not figures in 1987/8 dollar values. 741. Although, we do not doubt that, as a professional economist, Mr Gillett has a good general grasp of the distinction between constant and nominal dollars, we gained the impression, however, that he failed to bear it in mind when making readings from his trend line. When, as in the present case, regression analysis is based on data in nominal dollars, all readings from the trend line will themselves inevitably also be in nominal dollars. A trend figure of $2,354.00 in 1988/9 denominated dollars implies a figure in 1987/8 value dollars of $2,172.00 (using the 1988/9 deflator of 0.92284 shown in SF 216, page 163). Thus, Mr Gillett's trend line analysis, far from supporting his figure of $2,400.00 as a trend price, in fact supports a figure of $2,172.00 (in 1987/88 dollar values) which is even below the government's trend price of $2,200.00. 742. When challenged by counsel for the government to the effect that consistency with Mr Best's approach in the accounts required the trend price for rebars to be in the dollar values of 1987/88 and not 1988/89, Mr Gillett merely answered, "Well really that is Mr Best's problem and not mine." (Mr Gillett's Transcript, page 304, line 21). 743. Mr Best, when he came to give evidence, sought to down-play the significance of regression analysis, pointing out that the co-efficient of determination (i.e. the R-squared) of Mr Gillett's regression analysis was just over 74% (see Exh SF 40), which meant that only 74% of the trend price derivable from the line of best fit in Exh SF 34(c) was explicable on the basis of time. That implied that 26% of the movement in rebar prices stemmed from variables other than time. 744. The fact that the coefficient of determination was only 74% had not worried Mr Gillett who had described the trend line as, "not a bad fit". One has to bear in mind here, too, that it was Mr Gillett who had caused the trend analysis to be done, and that he was the one put forward by SFI as the expert on rebar prices. 745. Mr Li sided with Mr Gillett, agreeing that Mr Gillett's graph in Exh SFI 34(c) showed a statistically valid trend. 746. In turn, Mr Li did his own regression analysis based on rebar delivery prices from 1978/9 to 1987/8 (whereas Mr Gillett's was based on the A.W.P.I. from June 1971 to June 1988). That showed a trend price for 1987/88 of $2,035.00. The R-squared there was just under 65%, which Mr Li regarded as acceptable, but which, Mr Best said, was too low to show any statistically valid trend. We preferred Mr Li's view on this. (See Mr Li's report, 42/05B, page 102) 747. Mr Li's coefficient of determination got a boost to over 67% when he caused a fresh regression analysis to be done (see 42/05E page 19) to take account of an agreed upward revision of delivery prices. The closer the coefficient of determination is to 1, the higher the correlation between the variables of rebar price and time. 748. We ourselves have caused the parties to apply regression analysis to the delivery prices we have found as a fact for the 5 years 1983/4 to 1987/8. The coefficient of determination was 69.9%, and the line of best fit shows a trend price of $2,091, which is well below the figure we arrived at for delivery prices for the same period, inflated by the G.D.P. inflators issued by C. & S.D. We now set out the workings of that regression analysis and inflated averages : "REVISED PAGE 19 OF 42/05 EII SHUN FUNG IRONWORKS LIMITED CALCULATION OF TREND PRICE FOR STEEL BARS 1. REGRESSION ANALYSIS :
Projected selling price in 1987/88 (trend price) = $2,091 per metric ton 2. AVERAGE SELLING PRICE
Notes: 1. Selling prices from 1978/79 to 1981/82 are Shun Fung actual prices. 2. Selling prices from 1982/83 to 1987/88 are Hip Hing Delivery prices from Shiu Wing for standard length rebar inclusive of x 1.015. 3. Projected prices are based on the regression equation Rebar price = -4,326 + 72.92121 x year. 4. Inflation factors used in calculation of average bar price are GDP inflators. 749. Following Mr Li's approach, we have adopted the higher figure of $2,247 found by inflated averages as the basis of our trend price. Rounding up or down to the nearest $5, we find as a fact the trend price for rebar is $2,250.00. 750. Another complaint from Mr Best against the use of regression analysis was its unsuitability as a basis for predictions outside the period of the data. If that were correct, Mr Gillett needed saving from himself too, since he had attempted to say where the 1988/89 trend price was heading, although he only had data as far as August 1988. 751. In an endeavour to make his point, Mr Best explained that if, say, one attempted to extrapolate a figure for the 1960 trend price from data covering the 1971 to 1988 period, one finished up with a negative number. That is no doubt so, but we, nonetheless, felt that common-sense was clearly on Mr Li's side when he expressed the opinion that such data could by safely used for an immediately-following period such as 1988/89. 752. Leaving aside regression analysis, which only served to undermine Mr Gillett's trend price of $2,400.00, we now go on to consider whether there was any other evidence which might justify Mr Gillett's raising his trend price from $2,208.00 to $2,400.00. Mr Gillett was hard put to identify any specific happening in the external world between his First Report in February 1988 and Second Report in November 1988 which might justify his change of opinion about the trend price. He was, however, able to refer to recent events in Europe where, between April 1988 and September 1988, there had been massive increases in rebar prices, including an increase of 90% in Germany during that time. That does not seem to us significantly different in kind or degree from the increases he told us about in Japan and the United States in his February Report, which also included reference to the dramatic upsurge in prices in Hong Kong during the winter of 1987. 753. His Second Report increased the spot price in Hong Kong for February 1988 to $4,000.00, rather than the $3,400.00 referred to in the First Report. That, in our view, does not justify any increase in the trend price. He also mentioned how prices had hardened in Hong Kong since September, which was only two months prior to his giving evidence. Evidence of that sort did not, in our view, carry enough weight to justify changing a trend price which was intended to apply to a period of up to 10 years. 754. As we have already mentioned, the prediction in his First Report of an A.W.P.I. for 1987/8 turned out to be an over-estimate : instead of $2,536.00, the actual figure turned out to be $2,458.00. That took care of the position to the end of June 1988. Mr Carnwath, for the government, asked Mr Gillett whether anything had happened since the end of June 1988. To that, Mr Gillett had replied "No", and then went on to say how he had taken a more thorough-going look at the whole of the trend price over the past few years. That, then, led on to a consideration of exhibit SF34(c), with which we have already dealt. 755. The reality, as we see it, was that nothing had occurred in the external world to justify Mr Gillett's revising his trend price upwards. The only change that had taken place was in Mr Gillett's own mind. Whereas in his First Report, he had arrived at a figure on the basis of his intellect, we feel that the figure in his Second Report was based more on his emotions. The detachment which we are able to admire in his First Report had evaporated by the time of his Second Report and his actually giving testimony to the Tribunal. He has fallen into the classic trap for expert witnesses of becoming too closely identified with the client's cause. 756. We will now elaborate on the reasons identified by Mr Gillett for expecting a long term upward trend in rebar prices. 757. Basically, his observations amounted to saying that in 1986/7 and before, there had been excess capacity and over-production of rebars around the world. In such a situation, manufacturing rebars had not generally been profitable, since production costs exceeded sale prices. Now, however, there had been a reduction in capacity worldwide, and rebar-making was returning to profitability. 758. The impression we got from Mr Gillett's evidence was that he had already recognised the improving trend at the time of his First Report; nothing had happened after that to justify his raising of the rebar trend price to a higher level. The only fresh points introduced by his November Report (31/02) were the upsurge of prices in the E.E.C., particularly Germany, to which we have already made reference, and the existence of an international cartel, organised by the E.E.C., to keep steel prices up. 759. Mr Gillett's evidence on this international cartel was shadowy in the extreme, and we certainly did not get the impression that it has changed the situation of Hong Kong steel manufacturers being vulnerable to the dumping of cheap, steel in this territory. Such dumping imposes a ceiling on what local Hong Kong manufacturers can charge independently. 760. Although Mr Gillett in his evidence did make reference to a construction boom in the Pacific Rim area, we do not regard that as adding significantly to what he had already mentioned in his First Report about strong demand, and hence, high prices, for rebar in the United States, Japan and Hong Kong. 761. In the claimants' closing submissions, reference was made to the huge infrastructure projects planned for Hong Kong for the 1990's. The announcement by the Governor of plans for a new airport and harbour works did not get made until 1989, which was long after Mr Gillett finished his evidence. We do not consider the Governor's announcement as some sort of retroactive justification for Mr Gillett's increasing his trend price from $2,208 to $2,400. Although those infrastructure projects will obviously use a lot of rebar, the whole world is at liberty to flood Hong Kong with rebars, because of Hong Kong's comprehensive free-trade policies. We learnt from Mr Gillett that somewhere between 70% and 80% of the total consumption of rebars in Hong Kong during the years 1980 to 1987 was imported. 762. Another piece of information which came to light after Mr Gillett had given his evidence was,the A.W.P.I. of $2,682.00 for Financial Year 1988/89. Deflated back to 1987/88 dollar values, that becomes $2,475.00 which is slightly better than the A.W.P.I. for 1987/88 itself, which was $2,458.00. What relationship the A.W.P.I. for 1988/89 bears to 1988/89's delivery prices we do not know, and, on the view we have taken, it has been delivery prices which have been all-important to SFI for the purposes of the present case. Probably, it is reasonable to assume that 1988/89 delivery prices were at least as good as those for 1987/88 ($1,948), and probably even rather better. Thus, we accept that, in the No-Scheme-World, 1988/89 would have amounted to the continuation of an upward-trend in rebar delivery prices. 763. While Mr Gillett himself was realistically modest on the ability of anyone, including himself, to forecast rebar prices, powers were, nonetheless, claimed for him by Mr Best and Mr Widdicombe which went way beyond anything he had said himself. According to Mr Best, Mr Gillett's trend price represented average prices for rebar for the ten years starting 1988/9. According to Mr Widdicombe, Mr Gillett was able to look ten years ahead on rebar prices. Sensibly, Mr Gillett never contended he had those sorts of clairvoyant skills. 764. The truth of the matter is that standard length rebars are basically a commodity, not so different from, say, blocks of copper or bars of gold, when it comes to predicting future prices. Anyone who claims the ability to predict with some accuracy the price or even the trend in prices of a commodity over a period of years is deluding himself : there are just too many variables at workthroughout the world for anyone to have a significantly better than even chance of getting even the direction right. We only have to look at Mr Gillett's performance in trying to predict the prices of rebar and scrap eight months forward to be aware of how forlorn the hope is of coming up with anything like the correct prices. 765. We think that Mr Best was totally correct, when, in his First Report (33/01), paragraph 125, he referred to, "(a) the difficulty and uncertainty in predicting future prices generally; and (b) the particular difficulty and uncertainty in predicting the prices of steel bars and scrap, given the historical volatility of prices." He then went on, in paragraph 128, to adopt a numerical method for coming up with a trend price, namely, the average of the A.W.P.I. for the "high" year 1987/8 and the "low" year 1986/7. Further, he indicated in paragraph 129(c) that he had contemplated another numerical approach to prediction, namely, averaging over the most recent five year period. 766. What Mr Best said there was echoed by Mr Li in his report 42/05A at page 26 to the effect that, "it is not possible to predict future steel bar prices. The projection of a long-term trend price for steel bars is therefore wholly speculative". 767. It should be borne in mind, too, that it was Mr Gillett who had told Mr Best to average the "high" and "low" years to come up with a trend price for carrying forward in the Future Loss of Profit accounts. 768. After Mr Gillett and Mr Best had given their evidence, SFI propounded a new approach to ascertaining a trend figure. This approach is to be found in exhibit SFI216, page 249, paragraph 1 :
769. We cannot help feeling that SFI was there trying to make a virtue out of necessity, as Mr Gillett had performed so poorly on the numerical side, getting his predictions on the A.W.P.I. for rebar prices and the S.W.S. price for scrap wrong in his First Report, and having been shown to err in his own favour in mis-reading the trend line he caused to be prepared by way of regression analysis (See Exh SF 34(c) and SF 40). Against that background, paragraph 2, at page 249 of SF216 goes on as follows:
770. We do accept that Mr Gillett is an expert in steel markets, but reject the notion that his expertise extends to futurology. We think there is a high probability that if twenty steel market economists had given evidence in the case, we would have got twenty different answers. We do not feel that an answer based on subjective intuition is of much worth in the exercise to determine a trend price to be used in accounts for as long as ten years ahead. We have far more faith in an answer based on numerical and other analysis. Because we felt that, in the interval between his First and Second Reports, Mr Gillett had lapsed in the direction of partisanship, we were not inclined to accept his subjectively obtained figure of $2,400 in preference to his original figure of $2,208 which had the merit of being arrived at objectively by arithmetic. 771. Even if Mr Gillett's numerical approach had made more sense, and even if we had received the impression that Mr Gillett's evidence had been given with complete detachment, we would still not have regarded him as any better qualified than Mr Li to give an opinion on a trend price for rebars. With his background in accountancy, Mr Li was as well qualified as any steel economist to evaluate the evidence. 772. If, as we hold, regression analysis is a legitimate technique to assist in arriving at a trend price, we consider the correct data to use are delivery prices, rather than the A.W.P.I. to which SFI had recourse. Consistent with our approach to the Loss of Profits Claim for the Financial Years 1982/3 to 1987/8, we have used the SWS/HH delivery prices as the best indicator of what SFI's.delivery prices would have been in the "No-Scheme-World". Hence, our trend price for rebar is based on SWS/HH delivery prices. 773. Should it turn out that we have been wrong to follow the SWS/HH delivery prices rather than the A.W.P.I., then we consider that the A.W.P.I.to be adopted should have a four month time-lag. 774. We took note of the submission made on behalf of SFI that the trend price for rebars should reflect a long-term upward trend, and return to profitability for rebar manufacturing. In our view, that factor has been sufficiently taken into account when using the GDP inflators in Mr Li's inflated average approach. over the years, rebars have increased in price at a lower rate than general inflation. For example, over the period from 1981 to 1987, general inflation increased by 41%, whereas rebar prices, in nominal terms, increased by only 29%. Another indication of how rebar prices have lagged inflation is that even on the regression analysis done by Mr Gillett, based on the A.W.P.I. without lagging, the trend price for 1987/8 was $2,270, whereas an inflated average going back 12 years, as calculated by Mr Best in SFI 216 at page 160, was $2,574. If the rebar price had kept abreast with inflation, the trend price arising from that regression analysis should have been at least as high as the inflated average price. Those example are extracted from Mr Li's exhibit R106. 775. A suggestion that inflated averages of delivery prices should go back as much as 12 years was not spoken to by any of SFI's witnesses, and only got into evidence, indirectly, as part of SFI's so-called exhibit, SF216, at page 160. We incline to the view that the further one goes back, the greater the likelihood of distortion, because of the differing rates for rebar-price-inflation and general inflation. We think the five years suggested by Mr Li, (and implicitly accepted by Mr Best in his First Report (33/01) at paragraph 129(c)) is a reasonable period for present purposes. In effect, inflation-proofing the delivery prices adopted for SFI from 1983/4 onwards is a fair and reasonable way, in our opinion, to take account of any improvement in the outlook for rebar prices from 1987 onwards. 776. The reality is that neither Mr Li, Mr Gillett, nor ourselves have any precise way of making allowance for the increase in nominal rebar prices which was evident at the time of hearing this case. The future is unknowable. Whether the recently observable upward trend in nominal prices will turn out to be long-term, or merely a short period of remission in an otherwise downward trend, time alone will show. The best the Tribunal can do is to adopt a rough-and-ready approach in the face of an infinite number of imponderables and variables. We consider that the inflation-proofing approach we have adopted is the best and fairest available. 777. On the issue of the trend price, we sum up by noting that the Tribunal was offered the choice of Mr Li's massage-free, numerical approach, or Mr Gillett's intuitive method which, in some mystical way, arrived at an answer of $2,400. Based on the credibility of the witnesses and inherent probability, we prefer Mr Li's approach which produced a trend price of $2,250 (after rounding). Mr Best, despite being every bit as well qualified in accountancy as Mr Li, abdicated a direct role for himself in arriving at an appropriate figure, by putting forward the view, via his answers in Exh SF216 at page 249 in response to Mr Li's Exh R106 and R110, that this sort of forecasting was not the province of the accountant, and the Tribunal could only seek the answer in the figure given by a steel market economist, i.e., Mr Gillett. 778. Regarding that view, we do not for one second question Mr Best's honesty, integrity or sincerity - only his judgment. Away from the hot-house atmosphere of the present proceedings, we strongly doubt whether, after due reflection, Mr Best would continue to assert that only a steel market economist could come up with a trend figure, and that, in effect, a steel market economist is qualified to conjure a trend figure out of the air. 779. Loyalty is generally regarded as an admirable human trait, but not, however, when it is displayed by an expert witness who needs to cultivate detachment. 780. In allowing himself to be used to propogate the party line that only a steel market economist (viz., Mr Gillett) could give on authoritative trend price for rebar, we detect an erosion of the independence of mind manifested in Mr Best's First Report, and that helped diminish his effectiveness as a witness. SPECIAL LENGTHS 781. A standard length rebar is 12 metres. 782. Once SFI had got the concast machine working properly, it was no problem to make non-standard lengths. There was no extra expense in doing it. All that needed to be done was to inform the worker in charge of the shear, where billetts exited from the concast, to cut them to a particular length. SFI had a chart showing the length of billett required to roll a bar of a particular length and diameter. 783. For a building contractor, there were many advantages in being able to use rebars of a non-standard length. The contractor would be spared the trouble and expense of having to cut rebars to their required length on the site) and there was the advantage of not having any bar-ends to dispose of. Another illustration of the advantage of non-standard lengths to a contractor was given by Mr Gillett who explained how it might be worth a contractor's while to use two extra-length bars in a vertical reinforced column even if those two bars cost extra, rather than use three standard length bars. It is obviously less trouble for a contractor to have only two bars to handle, rather than three, particularly if the two bars have been manufactured precisely to the length he wants. Mr Mui of HH also gave an illustration of contractors preferring one bar of 10.5 metres to reinforce three floors of about 3 metres ceiling-height each plus their floor-slabs, rather than using a standard length bar, one and a half metres of which would need cutting off. 784. Contractors are prepared to pay a premium for non-standard lengths, because of the saving in costs they can make. 785. The attraction to a mini-mill in selling non-standard lengths is that, at no extra costs to the mini-mill; it gets paid more for its product. Besides that, a contractor who goes to a mini-mill for the non-standard lengths available there, is also likely to buy his standard lengths from the same place. The fact that an order for non-standard lengths is likely to be accompanied by an order for standard lengths is of no practical significance in the present case, since it was common ground that SFI, in the "No-Scheme-World", would have had no problem in selling all it could produce in any event. We are prepared to accept that, historically, in the Scheme-World, approximately 27% of the rebars SFI sold were special lengths, commanding a premium of, approximately, 6% above the price for standard lengths. In order to avoid a pretence to false accuracy, we opt for a round number in finding as a fact that SFI's earnings from rebar sales in the No-Scheme-World would have been boosted by 1 1/2% due to the premium for special lengths. 786. When Mr Gillett submitted his two Reports (31/01 and 31/2) and gave his evidence to the Tribunal, SFI sought an addition of only 1/2% to its overall earnings from rebar sales in recognition of the premium for special lengths. Mr Gillett had described that 1/2% as "most conservative". As we see the position, there is nothing in Mr Gillett's Reports, or the evidence from him or any of the other SFI witnesses, which estops SFI from claiming a higher amount than 1/2%, if the actual evidence justifies a higher figure. 787. One of SFI'S witnesses on rebar prices and special lengths was Mr Lam, formerly the Purchasing Manager of E. Man Ltd, the construction subsidiary of Henderson Land Ltd. From the evidence of Mr Lam and Mr Roy Leung, as well as from the evidence generally, we were left in no doubt that SFI would have earned considerably more than 1/2% from its special lengths in the "No-Scheme-World". 788. The very first contract in which SFI agreed to sell special lengths to a contractor at a premium was made in February 1979. From then on, SFI worked at developing a special niche for itself in special lengths. We have already mentioned how the concast machine facilitated SFI's making of special lengths. Other features of SFI's equipment conducive to the manufacture of special lengths were the three re-heating furnaces and the three rolling mills (although, as we show in our Section VI : PLANT AND MACHINERY, there are countervailing disadvantages to such an arrangement). A mini-mill such as S.W.S., which had just the one re-heating furnace and the one rolling mill, lacked SFI's flexibility to produce a variety of lengths of rebar. 789. We cogsider that by the early 1980s, the use of non-standard length rebar by the construction industry in Hong Kong was an idea whose time had come. That emerged strongly from the evidence of Mr Lam, as well as the evidence of Mr Mui, the manager of Hip Hing. Mr Lam, whom we regarded as a first rate witness, explained how, in the early 1980's, E. Man Ltd, which had, maybe, 25 or more major building contracts a year, set about standardizing the materials it used for building construction. Instead of buying materials separately for each contract, E. Man forecast the materials it would need for all its contracts on a global basis, and made its purchases accordingly. Mr Lam developed a computer software programme to assist E. Man in controlling its purchases of material. 790. One particular type of building material included in E. Man's programme was rebar, of both standard and non-standard lengths. Mr Lam pioneered the idea of using rebars of less than 12 metres in length, and for them, E. Man's only source of supply was SFI. Before Mr Lam's innovation, non-standard lengths were invariably longer than 12 metres. 791. An alternative possible source of supply for lengths in excess of 12 metres was S.W.S., and, very occasionally, steel stockists might have them. However, S.W.S. has not developed a niche in special lengths to anywhere near the extent SFI did. 792. Mr Lam, who had realized the cost-saving potential from using special lengths, whether of more or less than 12 metres, shared his knowledge with Sun Hung.Kai Properties Ltd, a major property developer in Hong Kong. E. Man Ltd. enjoyed a harmonious relationship with Sun Hung Kai Properties Ltd. We do not doubt that Sun Hung Kai Properties Ltd would have quickly come to appreciate the advantages of special lengths highlighted for them by Mr Lam's computerised programme. 793. Both Mr Lam and Mr Mui gave evidence to the effect that their reqirements of special lengths were generally somewhere in the region of between 10 - 20% of the total volume of rebars their companies used. We gained the impression, however, that they limited themselves to that range because there was no practical hope of getting more than that from the market, but, if more special lengths had been available, their range might have been higher. 794. In fact, the amount of special lengths ordered by Hip Hing from SFI Ltd during the period from September 1981 to June 1984 was 31.75% of all the rebars they took from SFI. The corresponding figure for E. Man around that time was 46%, and, on one particularly large development, (Telford Gardens in Ngau Tau Kok), 58% by volume of the rebars invoiced to E. Man by SFI were special lengths. 795. In the light of the evidence of what happened in the Scheme-World in the early 1980s, we are satisfied that SFI had identified a niche for itself as a manufacturer of special lengths, and, as already indicated, we regard an uplift of 1 1/2% over what SFI would have earned from standard lengths as reasonable. EARNINGS FOR THE YEAR 1988189 ITSELF 796. Whether the trend price for rebar is based on inflated averages or regression analysis of delivery prices, (as we have held), or whether it is based on the A.W.P.I., lagged or unlagged (as SFI maintains), that trend price comes into effect to represent SFI's earnings from the 1st July 1988, and onwards. Under the model agreed to by the parties, we regard it as unarguable to suggest that the point of commencement of the rebar trend price comes into effect on any day other than the Ist July 1988. 797. SFI has put forward a suggestion that if the Tribunal adopts a trend price based on a lagged A.W'.P.I. (which, for present purposes, we will assume to be of four months' duration), that trend price should not take effect until the beginning of the fifth month of Financial Year 1988/89, i.e. 1st November 1988, and for June till October 1988, the Tribunal should adopt the A.W.P.I. for the last four months of the Financial Year 1987/88, i.e. March to June 1988. Under that proposition, the a.W.P.I. for March 1988 becomes the figure to go into the accounts for the Financial Year 1988/89 as the July 1988 earnings, and the April 1988 A.W.P.I. becomes the August 1988 earnings, and so on. Such lagging of the A.W.P.I. can be found worked out in CCS section for Part I, Vol.2, page 9(a) of Appendix 4 at page 133 of that bundle. 798. The result of mixing the A.W.P.I. for March to June of Financial Year 1987/88 with SFI's suggested trend price of $2,400 from November 1988 to February 1989 results in a figure of $2,503 as SFI's earnings for Financial Year 1988/9. At page 67 of Mr Best's Report (33/06), we can see that figure of $2,503 incorporated into the Profit and Loss Account as the earnings from rebars for 1988/89. That table can also be found at page 9 of the Appendix 4 to which we. have referred in the last paragraph in relation to the A.W.P.I. lagged by four months. 799. What SFI is attempting to do with a blending of the trend price for part of 1988/89 with part of the A.W.P.I. for 1987/88 flies in the face of the agreement between the parties that the trend price for rebar takes over for the year 1988/89 and onwards. We think, too, that it flies in the face of SFI's own observation at para.4.5 of page 6 of Appendix 4 which is as follows :
800. We see no scope for selecting any date other than the 1st July 1988 for the adoption of the trend price. Even if the earnings based on the A.W.P.I. for March to June 1988 were in 1987/88 value dollars when SFI received them as earnings over the period July to October 1988 - And we do not think they would be 1987/88 value dollars being received in 1988/89 - we would not regard that as justification for postponing the commencement of the trend price until November 1988. 801. The circumstance that money from rebars happened to be in 1987/88 value dollars would be a necessary, but not a sufficient condition for that money to be treated as part of the 1988/89 earnings for the purposes of the accounts in the present case. Not only must the money be in 1987/88 constant dollars, but it must also be the trend price. The trend price, as we have already said, becomes the figure to count as earnings from 1st July 1988 onwards. 802. Under the model, the A.W.P.I. for March to June 1988 simply has no role to play in SFI's four-month lagged version, and the figures have to be disregarded. What the Tribunal needs to know under the model is SFI's earnings for the Financial Year 1987/8, and, on a four-month-lagged A.W.P.I. basis, those earning's comprise the A.W.P.I. from March 1987 to February 1988. What the A.W.P.I. was after February 1988 is of no concern for the purpose immediately in hand, since the model, which is concerned with earnings, already includes provision for earnings for Financial Year 1988/89, namely, the trend price. In this context, it has to be borne in mind that SFI's prices (represented by SWS or the A.W.P.I., as a surrogate) are only of relevance in so far as they translate into earnings, since it is earnings alone which feature in the Profit and Loss Account under the rubric."Prices Per Ton" "Sales" (See Mr Best's 33/06, pages 47 and 67). SFI's earnings under the model for Financial Year 1988/89 are the trend price. 803. We do not intend do labour the point, since we regard SFI's contention as devoid of merit. We will confine ourselves to pointing out just one anamoly which would occur if we accepted SFI's argument. The trend price for rebar and the trend price for scrap would be out of phase, with the scrap trend price taking effect from Ist July 1988 as contemplated by the agreement of the parties, whilst the rebar trend price would not become operative until lst November 1988. Instead of having all of SFI's outgoings and income in the 1988/89 Profit and Loss Accounts treated as being in 1987/88 dollar values with effect with 1st July 1988, there would be this remnant of the steel prices for the four months, March to June 1988, in 1987/88 nominal values at the time SFI got paid under its invoices for July to October 1988. 804. We do not intend to labour the point,, since we regard SFI's contention as devoid of merit. We will confine ourselves to pointing out just one anamoly which would occur if we accepted SFI's argument. The trend price for rebar and the trend price for scrap would be out of phase, with the scrap trend price taking effect from 1st July 1988 as contemplated by the agreement of the parties, whilst the rebar trend price would not become operative until 1st November 1988. Instead of having all of SFI's outgoings and income in the 1988/89 Profit and Loss Accounts treated as being in 1987/88 dollar values with effect with 1st July 1988, there would be this remnant of the steel prices for the four months, March to June 1988, in 1987/88 nominal values at the time SFI got paid under its invoices for July to October 1988. Table 1
February 1982 - January 1983
Table 2
March 1982 - February 1983
Average Wholesale Price
March 83 - February 84
Average Wholesale Price
March 84 - February 85
Average Wholesale Price
March 85 - February 86
Average Wholesale Price
March 86 - February 87
Average Wholesale Price
March 87 - February 88
Average Wholesale Price
| |||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||