Each Thursday we present a significant excerpt, usually from a recently published book or journal article. In every case the proper permissions have been obtained. If you are a publisher who would like to participate in this feature, please let us know via the site’s contact form.
Hilary Young, Associate Professor, UNB Fredericton Faculty of Law
Excerpt: Introduction, Results: Damages, and Conclusion
[Footnotes omitted. They can be found in the SSRN version via the link above]
This article presents the results of the first quantitative study of Canadian defamation actions. It examines the reported decisions between 1973-1983 and between 2003-2013 and includes a range of findings with regard to success rates, quantum of damages, proportion of corporate versus human plaintiffs, proportion of libel versus slander actions, proportion of cases involving journalism, proportion of cases involving new communications technologies, etc. It aims to contribute to debate about defamation law reform and to inform scholarly work in defamation law or in tort law and remedies more generally. Finally, this study should be of interest to practitioners who want to be better informed about the risks and rewards of defamation litigation.
The law of defamation aims to protect reputation while not unduly inhibiting free speech. It is an old tort that has changed relatively little over the centuries, despite changes in societal views about the importance of speech and despite radical changes in communications technology. In essence, a prima facie case is established if the defendant communicated about the plaintiff in a manner that would make a reasonable person think less of the plaintiff. The elements are often straightforwardly established and much of the free speech-protecting work of defamation law happens through defences, such as truth (justification), fair comment and qualified privilege.
Defamation is an unusual tort: it is strict liability (although the new defence of responsible communication is grounded in a lack of fault). Malice is presumed. Falsity is presumed. Damages are presumed. That is, the plaintiff need not prove actual injury to reputation in order to be entitled to damages. Damages are at large and are meant to both compensate the plaintiff and to vindicate her reputation. Unlike for personal injury claims, non-economic damages for defamation are not capped. Special damages are rarely sought or awarded. Defamation therefore differs from most torts, which require fault and require proof of injury for compensatory damages beyond the nominal.
In recent years, common law countries have been reforming the law of defamation. Australia and the United Kingdom have undertaken legislative reforms. In Canada, the Supreme Court has acknowledged that defamation law does not strike a proper balance between free speech and protecting reputation, and since 2008 it has been making incremental changes to defamation law to better protect expression. In addition, the Law Commission of Ontario has recently embarked on a defamation law reform project.
The argument for change tends to be based on principle (we are undervaluing free speech) or on anecdotal evidence (important speech is being chilled because of fear of litigation). This article provides, for the first time in Canada, a wide range of quantitative data about defamation actions that can inform the policy debate.
The study has a number of interesting results: it demonstrates that the average non-pecuniary damages award has more than doubled between the two periods studied – even when adjusted for inflation. It reveals that the percentage of corporate defamation claims (versus those brought only by human beings) is significant – about a third – but that the percentage has not increased greatly over the time periods examined. The study indicates that defamation claims resulted in liability significantly less often in the later period than in the earlier one. And cases involving new media (internet and email) publications resulted in liability more often than those involving other media of publication. The study shows that punitive damages are awarded to corporations more often than to human plaintiffs, and in higher amounts. It also shows that punitive damages were awarded in about a quarter of all defamation cases between 2003-13. These and other results are examined in detail below.
Before beginning the discussion of damages, it should be noted that there is at least one commercial product that will assess the quantum of damages for Canadian defamation cases. I intended to compare my results to those obtained in Westlaw Next Canada Defamation Quantum Service [the Service]. For a number of reasons, however, I did not. First, the cases the Service relies on are different than those that I rely on; only Westlaw cases are included in the Service. As a result, the Service found only 22 damages awards for January 1, 1973-December 31, 1982 (of which some were duplicates, as indicated below). My methodology uncovered 88 awards from 75 cases. Second, the methodology used to obtain data for the Service was sometimes difficult to discern (e.g. what counts as a “telecommunications” case?). Third, damages values are not adjusted for inflation. Fourth, the Service often counts the same case multiple times, even where the case involves only one award to one plaintiff. For example, it counts Vigna v Levant six times (that is, it counts it as six awards) for the purpose of the number of awards, average damages awards etc. Vigna is a trial decision with one damages award to one plaintiff. There were many examples of such duplicates. Thus, although the Service undoubtedly has utility, and Westlaw is working to eliminate duplicates and adjust damages for inflation, I consider the Service’s data for things like the number of awards, and average and median damages awards to be unreliable at this time.
One of the most interesting findings of this study is the large increase in damages awards over time. Each award was counted, so there is sometimes more than one per case.
The average total damages award in the 2003-13 cases ($61,662) is more than double (2.2 times) what it was between 1973-83 ($27,464) when adjusted for inflation. This is a statistically significant differences (p-value=0.000).
It is easy to imagine how general damages in particular could have increased significantly, since they are at large and reflect the value of losses without a market value: loss of reputation and pain and suffering. (Although total rather than general damages are referred to here, as we shall see below, they make up the lion’s share of total damages.) In Andrews, Justice Dickson referred to the quantification of non-pecuniary damages as “arbitrary”. Without an objective economic measure to ground quantification, it is easier for the quantum to vary and to change over time.
An increase in defamation damages awards is also consistent with David Potts’ observation that awards were significantly higher after 1995, when Hill v Church of Scientology upheld a $1.6 million award and rejected a damages cap in defamation.
Interestingly, in a study of British Columbia cases, David Gooderham suggests that most defamation awards are still relatively low, but that there are a few cases in which there are large awards of damages. Such an uneven distribution would mean that the average damages award does not reflect the kind of damages most plaintiffs can expect if they establish liability.
This claim can be tested by looking to the median rather than the average damages award. The median total award in the 1973-83 period was $12,671, which is less than half of the average damages award of $27,464. The median total award in 2003-2013 was $29,365 – almost half of the average award. In both periods, then, the median award is significantly less than the average, which means that most plaintiffs were awarded significantly less than the average. In fact, in the 2003-13 cases, 147/196 or 75% of the awards were below the average. A few plaintiffs obtained much higher damages awards, raising the average.
Note that the median in the later period is more than double that in the earlier period (132% higher). This means that it is not only the number of six figure awards that is higher. Awards at the lower end of the scale are higher too.
The chart below shows the number of awards, in each time period, within a range of dollar values. It demonstrates that there were many more (relatively) low value awards than high value ones in both periods.
The chart below is identical to the one above, except that the figures are not adjusted for inflation. It shows much lower damages awards for the 1973-83 cases, but even the 2003-13 cases are affected. For example, note that the highest damages award in the latter period was $750,000. The case is from 2003. Adjusted to 2013 dollars, the value of the award is almost $900,000. This chart demonstrates the importance of adjusting damages awards for inflation when comparing their value to that in other cases. The point is eloquently made in an article by Audra Ranalli and Bruce Ryder.
In order to determine whether the large increase in damages awards is evident within the studied periods, or whether it occurred largely in between 1983 and 2003, I determined the annual average defamation award. Especially in the earlier period, there were often only a few cases from which to determine an annual average, which helps explain the significant variation over time. A single very large or very small award could have a large effect on the average. This is the reason for the spike in averages in 2003: that was the year in which there was a $750,000 award (almost $900,000 in 2013 dollars).
Given the small number of cases per year, and the large amount of variation from year to year, no gradual increase is discernable. The average award is generally much higher in the years from 2003-2013 than from 1973-83, but there are exceptions.
Thus far I have been considering only total damages awards, but some interesting trends are worth noting with regard to general, aggravated, punitive and special damages.
In every case in which there was an award of damages, there was an award of general damages. This is expected given that general damages are presumed from a finding of liability. General damages constitute a majority of total damages. In 1973-83, they were $20,139/$27,464 or 73% of total damages. In 2003-13, they were $48,799/$61,662 or 79% of total damages. The average general damages award in 2003-13 is 2.4 times as high as that in 2003-13, which is similar to the size of the increase for total damages. The difference in the average general damages award between 1973-83 and 2003-13 is statistically significant (p-value). The difference in total damages between the time periods is therefore largely due to a difference in general damages awards.
Aggravated damages were seldom explicitly awarded in 1973-83 (in two cases only, out of 88 in which damages were awarded) but when they were awarded, the amount tended to be high: $55,351 and $27,539. In the 2003-2013 cases, the average aggravated damages award (again, only among cases in which aggravated damages were awarded) was lower – $30,560 – but that is effectively meaningless, given that there were only two awards in 1973-83.
Interestingly, however, in the later period there were many more aggravated damages awards relative to the total number of cases in which an award was made: 56/196. That is 29% compared to 2% of cases that had an aggravated damages award between 1973-83. Caution is warranted given the small number of cases in the 1973-83 data, but the data seem to suggest that it was much more common for courts to award aggravated damages in defamation cases between 2003-13 than between 1973-83.
This is interesting not only because of the difference, but because of the relatively high percentage of cases in the later period that had an award of aggravated damages: 29%.
The fact that aggravated damages are awarded in 29% of 2003-13 cases is arguably problematic. First, aggravated damages are controversial in an of themselves. The Ontario Law Reform Commission recommended abolishing them in 1991, although that call was never taken up. Raymond Brown also thinks that aggravated damages should be eliminated in defamation actions. The problem with aggravated damages relates to their tendency to overlap with general damages and punitive damages, creating the risk of overcompensation.
Second, the threshold for awarding aggravated damages is said to be quite high. Conduct justifying aggravated damages must be “reprehensible or outrageous”, “particularly high-handed or oppressive”, and “rub[bing] salt into the wound”. If aggravated damages are meant to involve such a high threshold of conduct, arguably we should not expect to see so many cases involving aggravated damages, especially given the availability of at large general damages awards and punitive damages awards.
The average punitive damages award among cases in which punitive damages were awarded was somewhat lower in 2003-13 ($14,002) than in 1973-83 ($17,376), despite the fact that total damages awards were much higher in the later period. Thus, the increase in total damages awards cannot be attributed to an increase in punitive damages.
The percentage of cases in which punitive damages were awarded was fairly constant: 20/88 (23%) in 1973-83 and 51/196 (26%) in 2003-13. However, given that defamation is strict liability and that punitive damages are relatively rare in tort, it is notable that a quarter of defamation damages awards in both periods include an award of punitive damages. A Cornell study of US cases found that punitive damages are awarded in 3-5% of cases (not just defamation cases) in which the plaintiff established liability. This figure reflects the exceptional nature of punitive damages. As with aggravated damages, the threshold for awarding punitive damages is high. “Punitive damages may be awarded in situations where the defendant’s misconduct is so malicious, oppressive and highhanded that it offends the court’s sense of decency.”
Yet Canadian courts imposed punitive damages in a quarter of defamation cases between 2003-13. Either courts are awarding punitive damages in situations that do not warrant such an award, or else defamation actions frequently involve malicious and oppressive conduct. If the former is the case, this is problematic. Such awards in defamation actions have an effect on free speech (both chilling and punishing it). This is why Australia has banned punitive/exemplary damages in defamation actions (although the United Kingdom rejected that approach).
For human beings, reputational injury often does not result in economic losses. In addition, current defamation law, which presumes more than nominal damages, provides plaintiffs with little incentive to prove special damages unless there is readily available evidence of such losses. Consequently, special damages are rarely sought or awarded. In fact, in none of the 1973-83 cases was there an award of special damages.
In the 2003-13 cases there were seven special damages awards out of 196 awards. Thus, 4% of damages awards in this period included an award for special damages. The highest award of special damages was $31,836 and the lowest (non-zero) award was $101. The total awarded in special damages was $93,244.
Before examining damages based on the type of plaintiff, there are a few other interesting things to note about damages awards generally. They are sometimes only nominal – $1 or $2 (the $2 figure in the chart below is a $1 award adjusted for inflation) despite the presumption of damages. In the 1973-83 cases, 12 awards out of 88 (14%) were under $1000 and 27/88 (31%) were under $5000. In the 2003-13 cases, six out of 196 awards (3%) were of $1000 or less, and 28 awards out of 196 (14%) that were $5000 or less. The proportion of awards under $1000 was significantly lower in 2003-13 than in 1973-83 (3% v. 14%) and the same is true of awards under $5000 (14% v. 31%). (All figures are adjusted for inflation.)
The maximum award is much higher in the 2003-13 cases than in the 1973-83 cases ($896,945 v. $276,754). It should be recalled that $896,945 is not the highest Canadian defamation award ever – just the highest in the 2003-13 reported cases. The highest reported defamation damages award was the $1.6M award in Hill v Church of Scientology, which would be worth $2,243,657 in 2013 dollars.) Further, there is a much greater percentage of awards over $100,000 in the later period than in the earlier – 17% versus 5%. This is a statistically significant difference (p-value = 0.000).
Damages awards by type of plaintiff
Average damages awards appear higher for corporate than human plaintiffs in 1973-83: $25,934 for human plaintiffs and $39,400 for corporate plaintiffs. However, that difference is not statistically significant (p-value = 0.611). Similarly, damages awards appear to be higher for corporate than human plaintiffs between 2003-2013 ($84,065 versus $53,358), but again, the difference is not statistically significant (p-value = 0.181).
The lack of a statistically significant difference is perhaps surprising, given that corporate reputations often have, at least in one sense, more value than human reputations. Then again, in defamation law, general damages for corporate plaintiffs should generally be low in the absence of proof of economic loss, and given that general damages are meant in large part to compensate for emotional and dignitary aspects of the injury to reputation, one might expect damages awarded to corporations not to be especially high.
Note that there is no significant change over time in the relative awards. That is, there is no evidence that corporations were awarded higher damages, relative to human plaintiffs, in the later period than in the earlier period.
Another significant and surprising finding with regard to damages to corporations is that in the 2003-13 data, punitive damages were awarded to corporations much more often, and in much higher amounts, than to human plaintiffs. In the 2003-13 data, punitive damages were awarded to human plaintiffs 20% of the time (in 28/143 cases) with an average award of $6,593. But they were awarded to corporate plaintiffs 43% of the time (in 23/53 cases) with an average award of $23,022. The difference between 20% and 43% is a statistically significant difference (p value = 0.002). In addition, the difference between $6,593 and $23,022 is statistically significant (p value = 0.004).
In the 1973-83 data, there were too few instances of punitive damages to draw meaningful comparisons. Punitive damages were awarded to human plaintiffs in 24% of cases and to corporate plaintiffs in 10% of cases, but the latter represents a single award.
Since whether punitive damages are awarded depends on the conduct of the defendant, the identity of the plaintiff should not matter. The difference in punitive damages to corporations versus natural persons in the 2003-13 cases therefore seems surprising. Were defendants behaving particularly egregiously when it comes to corporate reputation, or were courts more likely to find conduct that injures corporate reputation to be egregious – perhaps because corporations are less likely than humans to sue unless conduct is egregious? Further study is needed.
In the 2003-13 data, no significant difference was found between the frequency of aggravated damages awards to human (29% of cases) versus corporate plaintiffs (28% of cases). In the 1973-83 data there were too few cases from which to draw conclusions.
Damages awards by type of publication: journalism
The average damages awards for cases involving journalistic publications (newspaper, television or radio broadcast) versus non-journalism cases were compared. The average award for defamation involving journalism in the 1973-83 cases was $35,172. The average award in non-journalism cases was $15,361 (based on 34 awards in 29 cases).
The number of data points was not large, but the difference is significant at a level of α=0.05 (p-value = 0.016). Thus, the damages award was more than twice as high where the impugned publication was journalism than where it was not.
In the 2003-13 data, the quantum of damages awarded in cases involving journalism appears to be greater than in other cases ($62,169 versus $58,466), but the difference is not statistically significant (p-value = 0.799).
One possible reason for the disparity in the 1973-83 data is that journalistic publications often had broader reach than many non-journalistic publications (especially slanders) and there was therefore greater actual or potential reputational harm. Damages awards are meant to reflect the conduct of the defendant and the extent of reputational harm to the plaintiff. If this were the case, it might make sense that the disparity diminishes in the later period, in which anyone can publish to the world at large.
Damages awards by type of publication: new media
Damages awards in new media cases (internet and email) appear slightly lower than average: the average total damages award for all cases in the 2003-2013 period was $61,662, whereas the average damages award for a new media case was $57,386. However, the difference is not statistically significant (p-value = 0.716). Aggravated damages were also lower for new media cases: $5875 on average versus $8776 on average for all cases. The only type of damages that was higher for new media cases was punitive damages. The average for new media claims was $5552 ($16,656 considering only cases in which punitive damages were awarded). The average for all cases in 2003-13 was $3662 ($14,002 considering only cases in which punitive damages were awarded).
One might have expected higher damages in new media cases given that courts treat internet publication as an aggravating factor in quantifying damages. Yet the study suggests that there is no statistically significant difference.
To summarize, the average defamation award has more than doubled between the 1973-83 cases and the 2003-13 cases. The median, however, is much lower than the average in both time periods, meaning that most successful plaintiffs are awarded less than the average. Total and general damages are not significantly higher for corporate plaintiffs than for human ones. However, the number and quantum of punitive damages awards is significantly higher in the 2003-13 cases where the plaintiff was a corporation than where the plaintiff was a natural person. Between 1973 and 1983, total damages were much higher where a journalistic publication was involved, but that difference disappears in the 2003-13 cases. Damages in new media cases were broadly similar to those in cases where the medium of publication was not a new communications technology.
Damages Awards in default judgment cases
I hypothesized that damages awards would be higher when there was default judgment than when an action was defended. Defending the action should have the effect, on average, of lowering the damages award relative to leaving the action undefended. This hypothesis proved true in the 2003-13 data. (There were too few reported default judgment cases in the earlier data from which to draw reliable conclusions.)
In the 2003-13 data, default judgment was entered in 40 of the 196 cases (20%) in which liability was established. For these 40 cases, the average total damages award was $95,219. The average award in defended cases was $53,057. Recall that the overall average total damages award was $61,662. Thus, damages awards in undefended actions were 1.8 times those in defended actions. This is a statistically significant difference at a 5% level of significance (p-value = 0.031).
This finding is important when considering access to justice. Not only does not being able to afford to defend a defamation action affect the likelihood of being held liable, but it also affects the quantum of damages.
This quantitative study of Canadian defamation actions, the first of its kind, reveals a range of information that may be of interest to lawmakers, scholars and lawyers. Some of the study’s findings include: that general damages are significantly higher for corporate plaintiffs than for human plaintiffs; that the average damages award more than doubled between 1973-83 and 2003-13; that liability rates fell substantially between the periods studied and are quite low; and that damages awards used to be much higher where a journalistic publication was involved, but that seems no longer to be the case. Each of these findings, as well as the study’s other findings, can inform the debate about defamation law reform and provides a benchmark against which future changes can be measured.