Genetic Technologies Limited v Corson Grain Limited HC Gisborne CIV 2009-416-197

[2010] NZHC 1157

IN THE HIGH COURT OF NEW ZEALAND GISBORNE REGISTRY

CIV 2009-416-0197

BETWEEN  GENETIC TECHNOLOGIES LIMITED Plaintiff

ANDCORSON GRAIN LIMITED Defendant

Hearing:         28 June 2010

Counsel:         G Williams for Plaintiff

D J Sharp for Defendant

Judgment:      8 July 2010

JUDGMENT OF SIMON FRANCE J

Introduction

[1]      This case concerns the legitimacy of comparative advertising undertaken by the defendant.  The plaintiff says the advertising suggests that the defendant’s claims have scientific support which they do not.  It is said that the two brochures of the defendant which are in issue are contrary both to the Fair Trading Act 1986, and to the terms of a settlement contract entered into between the parties in 2005 following earlier similar litigation.

[2]      The present hearing involves a claim by the plaintiff for summary judgment on its proceedings.

Background

[3]      The parties are suppliers of seed products.   The seeds in issue are maize seeds.  The target audience are farmers and grain growers.

[4]      At issue are two brochures produced by the defendant:

a)        Corson Tender Leafy maize silage guide 2009/10;

b)        Corson maize grain guide 2009/10.

[5]      It seems that comparative advertising is a common feature of this area of commerce.   The recommended testing procedure is side by side growing over a number of sites and seasons.  Results will be relied upon in marketing.

[6]      In dispute is the level of statistical reliability that the results of the tests should have before being relied upon in such advertising.  There are in this case two facets to the dispute.

[7]      First, a term of the earlier settlement agreement between the same parties says that any claims made about products must be backed up by research results having a confidence level of 90% or greater.   The plaintiff says the defendant’s brochures breach this agreement.

[8]      Second, the plaintiff says that, putting the settlement agreement to one side, this level of confidence (greater than 90%) is anyway generally necessary, or comparative advertising will be deceptive and misleading.  The defendant disputes this and says that such a high standard of scientific reliability only favours the powerful who are resourced to do sufficient testing to achieve that level.   A requirement for results at that level will stifle competition and development.  One of the defendant’s experts says a 60% level of confidence should suffice.

[9]      A difficulty for the defendant is that many of the comparisons it publishes in the brochures have confidence levels lower than either figure; indeed, one had a

figure of 3.2%.  One can describe what that means in different ways, but effectively it says that if one repeated the same tests in similar conditions, there is only a 3% likelihood of the same outcome.

[10]     The defendant in publishing the comparative material did not expressly claim the test results had statistical significance.   Rather, unlike earlier years where it expressly disavowed statistical significance, the present brochures are silent on the topic.

The advertising material examined in more detail

[11]     The example described below is taken from the 2009/10 Tender leafy maize silage guide.  It is but one example from many in the brochure.

[12]     The defendant’s product in this example is CTL85, a hybrid maize seed.  The advertisement first identifies in written form some features of the product:

a)        early maturity hybrid;

b)        exceptional  yield  performance  for  cooler  season  environments

(Lower North Island and South Island);

c)        excellent staygreen;

d)preliminary data shows CTL85 has above average fibre digestibility, starch content and total energy;

e)recommended  plant  population  :  100,000  to  110,000  seeds  per hectare.

[13]     There  is  then  set  out  a  graph  which  compares  CTL85  to  competitors’ products in terms of the yield to be obtained from the plants.  The yield is assessed in terms of the total dry matter obtained per hectare.

[14]     The graph is in the form of an L-shape, with results for each product plotted on it.  The graph is headed:

Corson Silage Comparisons vs CTL85/Manawatu Taranaki Trials

The vertical left side line of the graph represents the yield difference.  The horizontal bottom line represents the dry matter % difference.

[15]     The defendant’s product is centred at the bisect of the two measures, and is given a value of 0.0%.  All the competitors’ products are located variously along the bottom line, but vertically all are below the CTL85 entry, thereby indicating that their left side figure, the key figure of dry matter yield, was less than the yield obtained from the CTL85 plants situated alongside.

[16]     Remembering that CTL85 is centred at 0.0, the competitors’ products are plotted at:

– 0.12 (so very close in yield to CTL85)

– 0.30

– 0.35

– 0.89

[17]     Despite  the  defendant’s  submissions  to  the  contrary,  in  my  view  it  is impossible to read this graph as being anything other than a claim that CTL85 outperformed its named opponents in these tests.

[18]     And it did.   The figures are accurate.   The issue is what is not said.   In particular the confidence levels for the tests (the degree of assurance one can have that similar tests will produce similar outcomes) are respectively:

11%

20.4%

34.2%

64.9%

[19]     These confidence level figures are the defendant’s own calculations.   It is reliability ratings at these low confidence levels that lead the plaintiff to say the graph is not only a breach of the settlement agreement, but also in itself amounts to misleading advertising.  As noted, the graph example analysed above is but one of

numerous similar graphs in each of the brochures.  One of the graphs on the next page of the brochure has confidence levels of 36.8, 18.1, 31.4, 37.2 and 3.2%.

[20]     Returning to the advertising concerning CTL85, the result between CTL85 and the nearest opponent (39F58) is next displayed in a block graph showing the actual quantities of dry matter yielded in the tests.  The figures are 22.82 tonnes for CTL85 compared to 22.70 tonnes for 39F58.  Below that block graph it is said:

Corson Silage Comparison

Manawatu Taranaki / Harvest Seasons 3

15 On-Farm Trials.

[21]     This description applies only to the single block graph comparison.   The number of seasons and trials is different for the other products featured in the main plot graph; for them the figures are 9 trials over 5 seasons, 11 trials over 3 seasons, and 9 trials over 4 seasons.  The plotted graph has the number of trials recorded in parenthesis next to the product name – e.g. 39G12(9) but not the number of seasons over which these 9 trials were done.

[22]     Finally on this example of the advertising, I note that in the trial  being discussed there were actually seven products tested against CTL85.  In addition to the four competitors’ products concerning which results were published, there were three others not referred to.  These three products all did better than CTL85.  The confidence  levels  for  these  three  tests  were  20.05,  21.19  and  65.07%.     In submissions Mr Sharp observed that results can be omitted from the advertising for legitimate reasons.   Whilst that is undoubtedly so, I observe that the graph on the following page which contains the test with a 3% confidence level, the same feature occurs of omitting all the better performers.  Further, in that example all the omitted tests had higher confidence levels than the published ones.

[23]     Finally by way of providing context to the plaintiff’s objections about this advertising,  it  is  fair  to  set  out  an  extract  from  the  equivalent  2007  brochure published by the defendant.  Concerning the test results that were published in the brochure, the defendant’s brochure first set out an explanation about the statistical significance of the test results.  The brochure contained the following table:

Number of Stars	Significance Level	Confidence Level	Conclusion
☼☼☼	Very High	≥ 99%	A very highly significant difference exists between the yields of the two hybrids.    Yield advantage can be claimed by the superior hybrid and can be selected with confidence.
☼☼	High	≥ 95%	A highly significant difference exists between the yields of the two hybrids. Yield advantage can be claimed by the superior  hybrid  and  can  be  selected with confidence.
☼	Modest	≥ 90%	A modest significant difference exists between the yields of the two hybrids. Although a yield advantage exists more trials are required to confirm hybrid superiority.  The hybrid with the yield advantage should probably be selected.
ns	Low or None	>90%	Little or no significant difference exists between the yields of the two hybrids. Yield advantage cannot be claimed by either hybrid.    The two hybrids compared  either  do  in  fact  have  a similar yield or in the case where a low number  of  comparisons  exist,  more trials are required to indicate whether a true difference in yield exists.   Hybrid selection should be based upon relevant traits.

[24]     Consistent with this, the published test results were all qualified by reference to this chart.  The second to last column of each table of results identified a yield advantage to the defendant’s products – e.g. 0.41, 0.50 etc.  The last column then set out the statistical significance of the result by use of either “ns” or 1, 2 or 3 stars.

The settlement agreement

[25]     Clause 3 of the settlement agreement contains definitions, including one for

“claim”:

“Claims”  include any statement or statements about Products including any statement relating to product performance and/or specificity for purpose of Products;

[26]     Clause 4 requires recitation in full, with the note that clause 4(b) is especially relied on by the plaintiffs:

4.        The parties will not:

(a)       distribute  any  brochure  or  other  Promotional  Materials which make any misleading or deceptive statement or statements about the performance characteristics of GTL’s and/or Corson’s Products or misleading or deceptive comparison or comparisons between them or make any such statements orally or by any other means;

(b)       make   any   claims   regarding   the   performance   of   their Products unless those claims are supported by verifiable scientific research  from  New Zealand  that  provides  statistically  significant results or credible non-partisan overseas research and, in particular, Corson will not make any of the statements or representations set out in Schedule 1 or any statement or representation that is in substance the same as those set out in Schedule 1.

The parties agree that claims about the silage quality traits or suitability of their Products for maize silage (and, in particular, the digestibility of grain or stover grown from their Products) will only be made if they can be supported by robust scientific evidence.

For the avoidance of doubt, for the purposes of this Settlement Agreement “statistically significant results” means result(s) having a confidence level of greater than or equal to 90.0%. (emphasis added)

[27]     Clause 5 is also relied upon.  It reads:

The parties agree that all Claims  made  in the future in  their respective brochures and Promotional Materials relating to Products will comply with the standards set out in Schedule 2 to this Settlement Agreement.

[28]     Schedule 2 which is referred to in clause 5 is lengthy, and it is not necessary to set it out.   As a whole it represents an agreement to observing fair advertising practices, and avoiding denigration of the other party.   It sets out predictable requirements for reliability of scientific results, and a process by which information supporting scientific claims is to be provided upon demand.

Decision

(a)      Breach of settlement agreement

[29]     The defendant has a considerable mountain to climb as regards the settlement agreement.  On its face the brochures appear to plainly be in breach of clause 4(b).

That clause requires claims regarding products to be supported by statistically significant results which in turn are defined as results carrying “a confidence level of greater than or equal to 90%”.

[30]   I prefer to address immediately the defendant’s position on this before considering the Fair Trading Act claims.

[31]     The essential argument advanced by the defendant is that clause 4(b) does not apply to the graphs because they are not “claims” as defined in the agreement. Rather,  the  graphs  are  comparative  material,  and  governed  by  clause 5  and Schedule 2.   It is argued that the term “claims” in clause 4(b) is limited to, and directed at, express claims such as “CTL85 has a yield advantage over the other four products”.    By contrast the graphs just provide information from which a reader might infer that conclusion.   It is submitted that as long as the brochure does not expressly say it,  but  leaves  it  to  inference,  there  is  no  “claim”  as  required  by clause 4(b).    Publishing  comparative  material  and  results  is  argued  not  to  be  a statement as that word is used in the definition of claim.

[32]     I do not regard the defendant’s interpretation of the contract as tenable.  The context in which the contract was formed was a desire to reach settlement over existing complaints about misuse of test results.   The is nothing to suggest that clause 4(b) should be read in such an ineffective way.   Clause 4(b) refers to any claims about a product, and the presentation of the tests results can only be seen as a claim about the performance of CTL85.

[33]     Mr Sharp sought to draw support from the use in clause 4(a) of the terms “deceptive statement” and “deceptive comparisons”.  He suggested that this showed that comparisons are different from statements.   Further, since “claim” is defined only  in  terms  of  “statements”,  clause 4(b)  which  refers  only  to  claims  must necessarily not cover this sort of comparison by graph.  However, in my view, when one reads clause 4(a) as a whole, the point is not valid.  The last part of clause 4(a) talks of “any such statements” in a way that clearly refers to all that has gone before

– i.e. both deceptive statements and deceptive comparisons.  Also, the reference at

the end of clause 4(a) to “orally or by any other means” makes no sense if the only claims caught by clause 4 are those expressed in the written or spoken word.

[34]     The present brochure represents a departure from the way the material was presented in 2007.  There the information was presented in a horizontal table form:

Hybrid	Harvest years	No. of Comparisons	Harvest DM%	Yield (tDM/ha)	Yield Advantage to CTL108 (tDM/ha)	Significance level
CTL108 3[   ]6	06, 07	18	31.3 29.5	25.82 25.49	0.33	ns
CTL108 33G26	01, 02, 03, 04, 05, 06, 07	60	33.6 32.6	25.84 25.16	0.68	ns
CTL108 34D71	05, 06, 07	37	33.1 34.8	24.55 23.58	0.97	☼☼
CTL108 34K77	01, 02, 03, 04, 05, 06, 07	54	33.0 35.2	25.60 24.78	0.82	☼☼
CTL 108 DK567	01, 02, 03, 05, 06, 07	35	33.7 36.4	25.16 24.18	0.98	ns

[35]     In  my  view  there  is  no  doubt  clause 4  applied  equally  to  this  form  of presentation.  The inference to be taken from how this material was presented is that the parties considered clause 4 applied to the table.  They were right to do so.  On the present occasion a different form of presentation of the same material is being used, but the applicability of clause 4 is unchanged.

[36]     Mr Sharp points out that if the current brochure breaches clause 4, so did

2007’s because statistically not significant results are still being used.  The point he makes is that acknowledging that the results have no statistical significance (as the last column does) cannot mean it does not breach clause 4.

[37]     I see the point, and tend to agree, but prefer not to determine it since it is not in issue before me.  The parties have, up till now, both freely used test results that have a confidence level less than 90%, but always by acknowledging the lack of

statistical confidence.   The brochure which I am considering does not follow that practice, and no one has complained about the 2007 brochure, so I leave the issue there.

[38]     Finally as regards the interpretation of clause 4, there is within it a reference to Schedule 1 of the contract.   That schedule sets out some statements which are agreed to be prohibited, of which two are:

6.That  CTL90  has  superior  emergence  and  seedling  vigour  when compared with competitors.

7.That CTL108 is the top-performing commercial hybrid in Corson silage trials across New Zealand over the last three harvests against all competitors.

These are prohibitions about specific comparisons, and make it plain that, contrary to the defendant’s submission, clause 4 applies to comparative advertising.

[39]     I  conclude  that  both  current  brochures  are  repeatedly  in  breach  of  the contract.  In Brochure One 83% of the test results which are used have a confidence level less than 90%.  In Brochure Two, 55.5% of the tests do likewise.  The extent of the breaches is therefore significant.

[40]     The defendant seeks to resist summary judgment by relying on the expert evidence  of  a  Mr Hardacre.     Mr Hardacre  deposes  that  the  defendant’s  own calculations of the confidence levels are flawed because they are too unsophisticated in their analysis.  If better analysis were done which had more regard to matters such as the timing or the seasons in which the trials were done, the confidence levels may, in his opinion, be better than currently claimed.

[41]     In my view this evidence is too speculative to be the basis on which summary judgment is declined.   First, the figures being used are those of the  defendant. Although the defendant is not necessarily to be bound by them, it is significant that it is their own calculation of the confidence levels of their own tests.  The defendant, obviously operating  on  the  assumption  they were  correct,  has  acknowledged  in correspondence to the plaintiff that many of its results are below the required 90% confidence level.

[42]     Second, it is not a case of a few tests only.  As noted, in the first brochure alone more than 80% of the cited test results breach the contract.  There is nothing in Mr Hardacre’s  evidence  that  leads  one  to believe that  recalculation  would  have outcomes anything like the scale needed to remedy the breaches.

[43]     Third,  it  is  significant  that  the  defendant  has  not  actually  done  the recalculations.   I have  no evidence as to what  is involved, whether the data is available, and how long it would take.  What is known is that the plaintiff’s evidence was filed in October 2009.  That evidence included what is in effect a compilation of the  defendant’s  data  to  identify  exactly  how  many  tests  failed  the  standard. Mr Hardacre’s  evidence  was  filed  in  May  which  is  obviously  quite  recent,  but overall there has been ample time since October 2009 to produce this evidence if it was to be relied upon.

[44]     Fourth, there are aspects of the affidavit that raise concerns.   Mr Hardacre consistently condemns the confidence level figures as being the plaintiff’s analysis, when they are in fact the defendant’s.   He also expresses the view that a 90% standard is scientifically unworkable, requiring “exceptional and in my experience unavailable data”.   Obviously on a summary judgment there is no opportunity to explore  this,  but  the  reality  is  that  both  the  parties  consistently  claim  to  have achieved such results.  Although the defendant has used many results with a lesser figure, it has also claimed to have reached it on many others.  I also find it hard to reconcile Mr Hardacre’s claims that the level is unachievable with his claim that re-analysis might bring many more of the tests up to that level.   The points seem irreconcilable.

[45]     For these reasons I decline to defer judgment on the contract breaches.  The defendant’s  brochures  are  plainly  in  breach  of  the  settlement  contract  and  the plaintiff is entitled to orders  requiring their withdrawal.   The defendant has no tenable defence.

(b)      Fair Trading Act

[46]     The discussion of the Fair Trading Act claims can be relatively brief.  At trial the defendant will, in my view, face a difficult task in defending the advertising.  But this is an application for summary judgment.  The claim of misleading advertising is made in relation to figures that are themselves accurate.  It is a claim based on the more difficult assertion that the advertiser was obligated to give more information.

[47]     The defendant has tendered evidence to say that the relevant market would not be misled, and would know the limits of the information being provided.  The defendant has also filed evidence which claims that a confidence level of 90% or higher is not a pre-requisite to valid comparative advertising in this area.

[48]     The task  of  the Court  on  a summary judgment  application  is  set  out in Krukziener v Hanover Finance Ltd [2008] NZCA 187, (2008) PRNZ 162 at [26]. It is there stated that the Court must be left without any real doubt or uncertainty. Needless to say that level of conviction must be sourced in the evidence filed.

[49]     Although troubled by the advertising, I do not consider that I can be that sure on a summary judgment application that the relevant market is likely to be deceived or would misunderstand the limits of the test results.  The parties both seem to think there is legitimacy in publishing statistically irrelevant results.   That being so, it needs to be explored what the effect is of expressly acknowledging, or not, the lack of statistical significance.

[50]     The defendant argued that the relevant market was better off if it had this information, even though the confidence levels of the results were so low.  I did not find it a convincing argument, particularly where the low confidence levels were not acknowledged, but I cannot say that the defendant should not be able to advance this proposition, and explain how it assists to defend a claim of unfair comparative advertising.

[51]   The plaintiff’s evidence does not attempt, at this stage, to establish by independent experts that publishing results with a confidence level lower than 90% is of itself misleading or likely to be.

[52]     Given  these  factors  I  consider  the  Court  is  obliged  to  decline  summary judgment in relation to the Fair Trading Act claims.  I have not ignored the fact that some of the confidence levels are so low as to suggest, without the assistance of expert evidence, that to rely on them in their bare form must be misleading.  But the authorities are clear that the advertising must be read as a whole and the 3% figure is not representative.  It is, for example, the only single digit result.  Likewise I have not ignored that the defendant’s expert suggests a confidence level of 60% might suffice, and that many of the results are below that as well.  Whilst this supports the plaintiff’s case, there is room for argument and the matter is not, in my view, suitable for summary judgment.

(c)      A third publication

[53]     The defendant sent what appears to be an e-newsletter to its clients.   The newsletter, amongst other things, noted that:

The recommended planting population of a Tender leafy hybrid is up to 10% less than comparable dual purpose hybrids.  This means you can plant less seeds per hectare, saving you money.

[54]     The plaintiff complains that this is a claim that a Tender leafy hybrid gives a

10% greater yield than competitors.   The defendant disagrees, and says it was intended to refer only to its own comparable hybrids.  It is poorly worded, to say the least, if that was the intention but the evidence is simply too sparse at the moment to address on summary judgment.  Further, it is unclear what relief is required if the defendant did not intend a comparative claim.   Since that was not their intention, presumably they will not make the same unintended claim in the future given that it has now been brought to their notice.

Conclusion

[55]     The application for summary judgment in relation to the Fair Trading Act claims is declined.

[56]     The application for summary judgment in relation to the breach of contract is granted.  The parties wished to confer on the form of orders.  To assist, I confirm my view is that the contract prohibits the parties from making claims, whether comparative or not, and whether such claims are express or by clear implication, as to the superior performance of their products unless the underlying testing has a 90% confidence level.  I base that almost exclusively on the wording of the contract, with contextual reliance only on why the contract came into existence.

[57]   I make no finding on whether publishing such results with an express acknowledgement of statistical non-significance complies with the settlement agreement.  That issue would require much fuller consideration of the factual matrix, and legally of what evidence might permissibly be considered.

[58]     I am not aware of any reason why costs should not follow the event in the normal way but counsel may file memoranda if necessary.

Simon France J

Solicitors:

D J Sharp, Burnard Bull & Co, PO Box 946, Gisborne, email: davi[email protected]

G Williams, Bell Gully, PO Box     , Auckland, email: garry[email protected]