D.1 The Arbitration
66 The Distribution Agreement between Castel and TCL was made on 29 December 2003, and varied in a meeting on 29 May 2007 ("the Variation Agreement"). It contained a clause which provided for arbitration of disputes to be held in Australia. No question as to the proper construction of the Distribution Agreement or the Variation Agreement arises in the applications to set the awards aside or in the application to enforce them. It is accordingly unnecessary to set out their terms.
67 Castel referred a dispute for arbitration on 25 July 2008 and the Tribunal was then duly constituted. There was an initial dispute as to the extent of the Tribunal's jurisdiction under the arbitration clause which was resolved in an application to the Supreme Court of Victoria: TCL Airconditioner (Zhongshan) Co Ltd v Castel Electronics Pty Ltd [2009] VSC 553.
68 Both parties participated fully in the arbitration and were represented by senior and junior counsel and solicitors. Procedural hearings leading to directions by the Tribunal were held in February, July and September 2010, and rulings were given on contested discovery issues. The substantive hearing was held in Melbourne for 10 days between 13 and 24 September 2010. Both parties' expert witnesses provided detailed reports, lay witnesses provided signed statements, and both were subject to cross-examination. Detailed opening and closing submissions were made by both parties.
69 It is unnecessary for me to set out all aspects of the arbitral hearing as no question arises as to the bulk of the 95 page Award. It is only in the approach taken by the Tribunal to the assessment of Castel's loss arising from TCL's sales of OEM products that TCL complains of breaches of the no evidence rule and the hearing rule.
70 Castel claimed that it was appointed as TCL's exclusive distributor in Australia for TCL-manufactured air conditioners, whether TCL-branded products or OEM products. It said that the Variation Agreement came about as a result of extensive complaints by it concerning, amongst other things, TCL's breaches of its obligations not to sell air conditioners that competed with the TCL-branded products that Castel had the exclusive right to sell. Castel contended that in breach of its obligations TCL supplied a large volume of OEM products under a variety of brand names to purchasers in Australia.
71 TCL accepted that it had sold 52,878 OEM products over the period of the Distribution Agreement and up to 31 December 2008 but denied that this was in breach of the agreement. It contended that the distributorship and exclusivity granted to Castel was confined to TCL-branded products and did not relate to OEM products. That is, TCL's obligation to Castel with regard to selling competing air conditioners in Australia was limited to an obligation not to sell TCL-branded products other than through Castel.
72 Contrary to TCL's submissions, the Tribunal determined that the Distribution Agreement gave Castel an exclusive right to sell all TCL-manufactured products in Australia. It determined that TCL's sales of OEM products over the period from 1 January 2004 to 31 December 2008 were made in breach of the agreement, and that Castel was entitled to damages. These findings are not the subject of complaint before me.
73 Having decided that TCL's sales of 52,878 OEM products were a breach of the exclusivity provisions of the Distribution Agreement it fell to the Tribunal to determine what loss Castel had suffered by reason of the unauthorised sales. Castel complained of very significant lost sales of TCL-branded products (and significant lost profits) by reason of the sales of the OEM products. TCL contended that Castel had suffered no loss from the sales of the OEM products, although leading expert evidence that the maximum lost sales were in the order of 7.4% of the sales of OEM products.
74 The Tribunal considered that the appropriate way to approach assessment of Castel's loss was by proceeding in three stages:
(a) estimating the additional volumes of TCL-branded products which would have been sold by Castel had TCL not sold OEM products in Australia in the relevant period;
(b) determining the prices at which those additional volumes would have been sold and the resulting gross profit margin; and
(c) calculating the additional costs which would have been associated with those higher sales volumes.
The Tribunal's three stage approach is not the subject of dispute in these proceedings. TCL complains only that, in the first stage of this process - that is, in estimating the additional sales of TCL-branded products that Castel would have made if the sales of the OEM products had not been made - the Tribunal failed to accord it natural justice.
75 In support of its claim of lost sales Castel called as an expert witness Mr Peter Acton, a financial and management consultant, as well as calling evidence from its Managing Director, Mr Michael Kwong, and from its Queensland manager, Mr Trevor Francis.
76 TCL called as an expert witness the economist, Mr Phillip Williams, to provide an opinion as to Castel's lost sales arising from the sale of the OEM products. It also called evidence from its General Manager, Overseas Business Division, Mr Shi Weiyi (Mr David Shi), and its Sales Director, Mr Frank Wang. An accountant, Mr Darryn Hockley, was also called by TCL as an expert witness but his evidence is not relevant to these proceedings.
77 One of the issues affecting the Tribunal's determination of Castel's lost sales of TCL-branded products was the extent to which the OEM products were directly competitive with and substitutable for them. Mr Williams discussed horizontal and vertical differentiation of products at paragraphs 32-35 of his report. He said:
32 The difference between the two kinds of product differentiation stems from the number of buyers who rank one product over another. Products are said to be horizontally differentiated if buyers differ in their rankings of the products so that, if all products were available at the same price, not all buyers would demand the same products. By contrast, products are said to be vertically differentiated if buyers have the same ranking of the products so that, if all products were available at the same price, all buyers would demand the same (most highly-ranked) product. As Tirole notes, one reason for vertical differentiation is difference in quality.
33 Consider the example of domestic motor vehicles. These products are differentiated because some products are sold at higher prices than others. The patterns of consumer preferences among motor vehicles involve elements of both horizontal and vertical product differentiation. Elements of horizontal differentiation seem to be matters such as: (i) seating capacity; (ii) engine size; (iii) four-door or two-door; (iv) hardtop or soft top; and so on. These are elements on which potential consumers have different preferences; that is, even if all cars were available at the same price, consumers would wish to purchase different cars depending on the extent to which the characteristics of the cars corresponded to their preferences over attributes such as these.
34 By contrast, almost all consumers would prefer: (i) a new car over a second-hand car; (ii) a car with perfect paint work over a car with defective paint work; and (iii) a car manufactured by a maker with a reputation for high quality engineering over one manufactured by a maker with a reputation for shoddy design and manufacture. That is, these are elements on which consumers will have similar preferences; that is, if all cars were available at the same price, consumers would choose those that were new, had perfect paintwork and were manufactured by a maker with a reputation for high-quality design and manufacture. That is, all consumers will prefer higher quality over poorer quality.
35 Economists often speak of products being located in 'product space'. Product space may have both vertical and horizontal dimensions. The closer one product is to another in product space, the more likely a consumer is to substitute one product for another if the price of the first product were to rise or if it were to become unavailable.
78 Mr Acton did not claim expertise as to the substitutability of OEM products for TCL-branded products. He relied on the detailed evidence of Mr Kwong and Mr Francis who said that the OEM products had a high degree of similarity with the TCL-branded products (sometimes appearing identical and having the same service brochure, sometimes being advertised as in fact being manufactured by TCL). Mr Francis also gave evidence of some retailers who had switched to OEM products from TCL-branded products. Mr Acton's evidence based on this was that each sale of an OEM product meant that Castel lost a potential sale of a TCL-branded product. That is, the substitution ratio of OEM products for TCL-branded products was 1:1 or 100%.
79 During the arbitration TCL challenged the expertise and qualifications of Mr Acton to give expert evidence on the topics contained in his report. While the Tribunal received Mr Acton's evidence, it ultimately said in its reasons that although he was an honest witness his report failed to meet the criteria for an expert's opinion and no weight could be placed upon his evidence as an expert. Nevertheless, the Tribunal considered that it was still able to have regard to some parts of Mr Acton's evidence as evidence of a fact. It considered that his evidence remained of assistance as a framework and a methodology for the calculation of Castel's losses because the methodologies were either not challenged, or if challenged deserved to survive that challenge. I note in passing that, while TCL raised concerns about the number of references by the Tribunal to Mr Acton's evidence in the Award, nothing of significance in the result turned on his evidence.
80 It was common ground that both the TCL-branded products and the OEM products occupied a product space at the bottom of the Australian market. In his evidence Mr Kwong identified five different levels in the market and the brands which occupied those levels as:
Level 1 - Daikin;
Level 2 - Panasonic, Toshiba, Fujitsu and Mitsubishi;
Level 3 - Samsung, LG and Kelvinator;
Level 4 - Midea, TCL, Gree (although Midea was said to be at the top of this level and TCL at the bottom);
Level 5 - all other Chinese brands (including the OEM products).
Mr Williams did not quarrel with these suggested levels.
81 Mr Williams rejected as unreasonable the proposition advanced by Mr Acton that 100% of the sales of OEM products were diversions from potential sales of TCL-branded products. He said that his "primary reason for this opinion is that TCL's products were not likely to be directly competing with its OEM products and other products within the lowest-priced segment of the air conditioning market."
82 There was no attack on the expertise or independence of Mr Williams. He was however vigourously cross examined as to the approach he took to the substitutability of OEM products and TCL-branded products and the estimation of Castel's lost sales, and Castel submitted to the Tribunal that his estimation of Castel's lost sales was far too low and some of his opinion was abstract theorising.
83 Importantly, Mr Williams conceded an absence of data as to the degree of substitutability between OEM products and TCL-branded products. He noted:
56 …air conditioning units are a highly differentiated product. This means that the extent to which one brand is a substitute for another is a matter of degree. I have made efforts to discover data that would enable me to quantify the degree of substitutability between the OEM's supplied by TCL and the TCL-branded product. I have been unable to discover such a source of data.
(Emphasis added.)
In my view, the absence of such data meant that Mr Williams' estimate of the level of substitutability was unavoidably less reliable.
84 In a joint statement with Mr Acton arising from an expert conclave, Mr Williams accepted that the estimate of Castel's loss was made more difficult by the absence of appropriate data. The joint statement said:
1. As a matter of principle, the lost sales can be divided into two categories. The first is that end-users who bought TCL-OEM product would otherwise have bought TCL-branded product if there had been no TCL-OEM product. The second is that end-users who bought non-TCL product would otherwise have bought TCL-branded product.
2. In order to assess the magnitude of the first category, it would have been useful to see:
(i) point of sale data to indicate how end-users switched purchases as a result of changes in relative prices; and/or
(ii) customer survey or focus group data as to the decision criteria and product perceptions of end-users.
3. The only numerical data that seem to be available to assess the first category, relate to total sales of different products.
4. The problem with drawing inferences about the first category of lost sales from aggregate sales data is that we do not know whether the field of substitution was narrow or wide. At one extreme, all products may be in the field in which case the loss of quantity corresponds to the first Williams alternative. At the other extreme, the only other product in the field may be TCL-branded product in which case, the loss of quantity is the number of TCL-OEM sales.
5. The second category of lost sales is that end-users who bought non-TCL product would otherwise have bought TCL-branded product. This may have occurred if retailers became less willing to promote Castel products or if end-users became less willing to buy a Castel-distributed products because of bad experiences or reputation. We have seen no data that would enable us to quantify this effect.
(Emphasis added.)
85 Mr Williams offered an opinion as to how a rational business would behave in the circumstances, stating:
57 Nevertheless, I can comment on what a profit-maximising enterprise would have intended. No profit-maximising enterprise would deliberately position versions on which it was earning a lower margin to be close substitutes for versions on which it was earning a higher margin - because to do so would cause it to lose substantial quantities of higher-margin sales in return for lower-margin sales. If TCL were attempting to maximise its products, it would have attempted to position its OEM versions such that they were not seen as close substitutes for its TCL-branded product.
86 Mr Williams considered that "a key distinction between brands at level 5 and those at level 4 and above seems to have been the benefit of assured after-sales service. That is, a key element in a consumer's assessment of quality seems to have been the assurance of after-sales service." Castel offered after-sales service of the TCL-branded products and such service was not offered for the OEM products.
87 However, the Tribunal did not accept that TCL behaved in a rational economic fashion, in that it found that TCL took no active steps to protect its TCL-branded products from OEM products. The Tribunal also considered that the existence of after sales service was a less important product differentiator because retail sales staff would likely urge customers towards the product which offered the best profit margin for the retailer, regardless of whether after-sales service was provided.
88 In relation to the impact of the sales of OEM products the Tribunal noted:
Evidence of OEM sales and impact
196 The evidence on which Castel relied for the extent and effect of OEM products was the evidence of Mr Kwong and Mr Francis. Mr Francis in particular gave detailed evidence of the various OEM brands which he inspected and how we formed the view that they were TCL-manufactured products. In some cases they appeared identical, in one case he recognised the service brochure as one that he himself had written for a TCL-branded product. Mr Francis also gave evidence of how the presence of the OEM brands affected Castel's sales. He gave examples of accounts that switched to the OEM brand on the basis that it was TCL-made and cheaper or which reduced their purchases of Castel's products in favour of OEM brands. Castel also produced a newspaper advertisement that stated that a particular OEM brand was manufactured by TCL.
197 Mr Kwong's evidence was that the presence of TCL's OEM products at the low end of the market affected the positioning of the TCL brand, which was being extensively promoted by Castel and which had positive associations through its anti-allergy certifications.
198 Castel relied on that evidence as demonstrating that the presence of TCL's OEM products was a factor which affected whether wholesalers bought goods from Castel and whether individual retail purchasers chose a TCL-branded product or another product.
199 TCL submitted that the evidence of the impact of OEM products was wholly unsatisfactory and could not support the conclusion that the presence of TCL-made OEM products had had any effect on Castel sales.
89 In relation to the degree of substitutability the Tribunal noted:
Substitutability and product differentiation
205 Another contested issue that affects the determination of Castel's lost sales was the extent to which the OEM products were directly competitive with TCL-branded products. It was common ground between the parties that the air conditioning market is highly differentiated. Mr Williams gave a helpful summary definition of the two forms of differentiation: vertical differentiation means differences in the quality of the product, and horizontal differentiation means differences in the physical characteristics of units while they may be identical in their power and functioning. There is a range of styles and types of units available and customers make choices based on their particular needs.
206 Castel's claim was that the TCL OEM products were marketed in the same areas of the market, and had the same essential features, as the TCL-branded products so that they had a direct impact on whether TCL-branded products were bought. Mr Francis's evidence was that he inspected a range of units and found them similar or identical to models which Castel was selling. He also gave evidence that they were perceived by number of Castel customers as so similar to the TCL-branded products that they were an acceptable and cheaper substitute for them.
207 Mr Acton did not purport to give expert evidence as to the substitutability of OEM products for branded products. He made his calculations on the basis of Castel's evidence that there was a high degree of similarity.
208 TCL produced evidence in the form of pictures that compared TCL-branded units to OEM units. It is submitted on the basis of that evidence that the two types of units were insufficiently similar to be true substitutes. This evidence was somewhat undermined by other evidence from Mr Shi that one reason to make OEM products was that spread the product and development costs which would otherwise be wholly attributable to the TCL-branded products. Mr Zhang also gave evidence that TCL-branded products and OEM products were in effect the same units, with different trademarks. Castel submitted that the differences were purely a matter of form and that in functional matters the TCL-branded units and the OEM units were close substitutes for each other.
209 TCL's expert Mr Williams proceeded on the basis that TCL would not have been likely to cannibalise its own sales and TCL was therefore engaged in 'versioning' - the sale of different versions of its products to different sections of the market - so that TCL's OEM products would not have directly competed with TCL-branded products because they would have been sold in different places. Castel submitted that TCL's own evidence showed that this was not the case and that TCL took no interest in where and how its OEM products were sold. While there is no evidence that OEM product and a TCL branded product were ever side by side in a store, or even in adjoining stores, we accept that TCL was not taking active steps to protect its TCL-branded products from OEM products.
…
211 One key element of the TCL-branded products that was said to differ from the OEM products was the after sales service offered by Castel. Level 4 brands tended to offer this service where level 5 brands did not. Mr Williams' evidence was that this was an indicator of a vertical differential; customer services are an important element of quality and a reason for customers to prefer a higher level brand such as TCL. Mr Williams saw this as a sign that the OEM products would not be directly competitive with the TCL-branded products which offered that degree of support. He agreed in cross examination that sales staff have a high degree of influence on what a customer buys and that they would urge customers towards products which offered the best profit margin for them, regardless of after-sales service.
212 Having considered all of the evidence the Tribunal does not propose, and nor was it invited, to consider model by model whether there was direct substitutability between TCL-branded products and OEM products. It accepts the evidence of Mr Francis as to his observations and the perceptions of customers as conveyed to him. It accepts that for sections of Castel's actual and potential customer base the OEM products were regarded as just as good and led those customers to choose the OEM products instead of the TCL-branded products.
213 At the lower end of the market where the TCL brand and the OEM brands were competing it appears that functionality rather than form was the important issue. The TCL-branded products and the OEM branded products served in the similar part of the market and offered the same types of units.
214 The Tribunal cannot find as a fact that every OEM units sold was directly substitutable for a designated TCL-branded unit. However for the purposes of estimating the loss of sales to Castel brought about by the presence of TCL OEM products we proceed on the basis that the OEM products were sufficiently similar to be a direct competitor of, and replacement for, a line of TCL-branded products, at least insofar as they were perceived in that way by the customer base. It is important to note that Castel did not sell directly to end-users but rather to retailers who bought in bulk and who were well informed about the market and about the available brands.
(Emphasis added.)
90 Mr Williams considered three alternative approaches to the estimation of the lost sales of TCL-branded products. The first alternative was to approach the question on the basis that OEM product sales took away sales from all other brands including TCL-branded products, at an equal rate. The second alternative was to approach the question on the basis that OEM product sales took away sales from all other air conditioners that had similar physical characteristics and were priced at level 5 of the market, at an equal rate. Mr Williams rejected both of these approaches.
91 He considered the better approach was to assume that sales of OEM products were drawn equally from sales of all other air conditioners that had similar physical characteristics and were priced at levels 4 and 5 of the market. That is, he made his estimation of Castel's lost sales on the assumption that if the unauthorised sales of OEM products were not made, the consumers that made those purchases would have purchased other brands of air conditioner at levels 4 and 5 of the market in accordance with the market share of each brand. Mr Williams accepted that this approach was based on the implicit assumption that the sale of an OEM air conditioner is equally as likely to be at the expense of the sale of a Level 4 as a level 5 air-conditioner.
92 In relation to this assumption Mr Williams conceded at paragraph 94 of his report that:
The evidence I have examined as to whether this assumption is appropriate is inconclusive. On the one hand, there is evidence of vertical product differentiation in the manufacturing market…and this differentiation is reinforced by vertical differentiation in retailing markets…On the other hand, there is evidence from Mr Kwong that sales of TCL-manufactured OEM products damaged sales of TCL-branded product. My assumption of diversion to all brands within levels 4 and 5 in proportion to market shares might be seen as a compromise between these two kinds of evidence.
(Emphasis added.)
Mr Williams further indicated his uncertainty as to his assumption by stating in a footnote to his report: "This is the best answer I am able to give to question 8 in my letter of instructions." Question 8 had enquired:
Would end consumers switch from an OEM brand (for example, Conia) to a recognised brand such as the TCL Brand if they knew it was manufactured by the same company? Even if there is a distinct price differential between the cheaper OEM Brand and the TCL Brand?
93 Mr Williams then estimated the market shares of each brand of air conditioner within levels 4 and 5 of the market. Having done so he re-estimated the market shares assuming that the sales of OEM products made by TCL were redistributed to each other brand within levels 4 and 5 in accordance with their market shares.
94 In making his estimate of the Castel's lost sales Mr Williams relied on historical data (provided by a company named GfK) of Australian sales of different brands of air conditioners by major retailers ("GfK data"). Unfortunately, the data was incomplete in that
(a) it related only to major retailers whereas Castel sold to a variety of retailers including small retailers;
(b) it covered only 54% of the total market for air conditioners in Australia which meant that any conclusion based on the data was based on information relating to just over half of the market;
(c) it did not refer to any sales of TCL-branded products for the years 2004 to 2007 which was the majority of the relevant period; and
(d) it only referred to sales of TCL-branded products for a seven month period in 2008, at a time when the relationship between TCL and Castel had broken down - which meant it was unlikely that sales in this period reflected the normal market share of TCL-branded products.
95 Mr Williams was advised by GfK that although TCL-branded products did not appear in its database through much of the period, sales of TCL-branded products were recorded in a generic category named "Tradebrand". In that category sales of TCL-branded products were aggregated with sales of other air-conditioners and there was usually no separate itemisation of sales of TCL-branded products. GfK also advised that the Tradebrand category is one that was exclusively, or close to exclusively, sold in one particular retailer. In the relevant period 35.1% of the sales of air conditioners were recorded within the Tradebrand category, which meant it was a significant category.
96 As the evidence is that TCL-branded products were sold in a variety of retailers rather than in one particular large retailer Castel argues that the extent to which the Tradebrand category captured Castel's sales of TCL-branded products is unclear. I accept this. In my view (and as found by the Tribunal) the uncertainty as to this significant category operated to reduce the reliability of Mr Williams' estimate of lost sales based on the data. The other shortcomings in the Gfk data operated in the same way.
97 Because of the deficiencies in the GfK data Mr Williams found it necessary to make what he called "three important adjustments" before offering his opinion. First, because the GfK data did not include TCL-branded products except under the Tradebrand category, Mr Williams incorporated estimates of the sales of TCL-branded products based upon sales data provided by Mr Shi of TCL. Second, because the sales of OEM products recorded in the GfK data were generally larger than TCL said it had made, he used sales data provided by Mr Shi to adjust the market shares of the OEM products. A third adjustment to market shares was necessary because some of the level 4 and 5 brands specified in GfK's total air conditioner sales data did not appear in the information on air conditioner types which recorded the brand of the different types of air conditioners. To my mind, the requirement for these adjustments again illustrates that there were real questions as to the reliability of the data and as to any conclusions based on it.
98 After Mr Williams' adjustments were made he estimated that in the absence of sales of OEM products by TCL, Castel could have expected to pick up a maximum of 7.3 % of the foregone sales as extra sales of TCL-branded products. By a process which was not made clear (but which was probably just a rounding up) this assessment was treated before me as being 7.4%.
99 As the Tribunal noted at paragraphs 220 to 222:
220 …There were no independent figures available for the number of OEM sales by TCL or for the total sales of air conditioners in Australia. Mr Williams gave evidence that he had sought and obtained certain data from GfK on sales of air conditioner brands by major retailers in Australia. The data represented 54% of sales and referred to a number of brands for which TCL had manufactured OEM products. It did not specify who had manufactured those brands and for all but a small period in 2008 the data did not refer to TCL branded products, or to many of the OEM brands, other than in a catchall Trade brands category.
…
222 One significant difficulty faced by Mr Williams was that, as we have noted, in the GfK data on which he relied for his analysis the TCL brand was not expressly named. He gave evidence that he had made enquiries about whether the "Trade brands" category could be disaggregated to reveal precise TCL figures, but that those enquiries had not been successful. He agreed that this was an unsatisfactory outcome but said it was the best data that he had found.
(Emphasis added.)
100 The Tribunal went on to make the three findings that TCL complains about. While I will detail these three findings later, it suffices for the present to note that the findings meant that the Tribunal found that the total of Castel's lost sales was 22.5% of the sales of the OEM products, rather than the 7.4% proposed by Mr Williams.
101 TCL says that it was wrong for the Tribunal determine that Castel's lost sales were 22.5% of the OEM products sales, but it does not complain about the calculations that followed that determination. That is, it is not controversial that if Castel's sales of TCL-branded products over the 5 year period to 31 December 2008 had been decreased by 22.5% of the sales of OEM products, it had lost the opportunity to sell a further 11,708 air-conditioners. It was also uncontroversial that Castel's lost profit from such lost sales was $2,198,656. The Tribunal awarded this sum to Castel.
102 The Tribunal's assessment of Castel's loss was well short of the $12,315,233 that it had claimed for that period based on a 1:1 substitution ratio, but well more than the loss if the award was based on Mr Williams' estimate of 7.4%.