One of the most consistent themes of the fashionable cause of libel reform is that libel actions are vastly too expensive both for claimants and for defendants. We are confidently informed that libel costs “regularly reach hundreds of thousands of pounds”Footnote 1 and that they are “out of control”.Footnote 2 The issue has been taken up by the Commons Culture, Media and Sport CommitteeFootnote 3, a ‘Libel Working Group’ set up by the Ministry of Justice under the previous governmentFootnote 4, and, most notably, by the pressure group the Campaign for Libel Reform, sponsored by Index on Censorship, English Pen and Sense about Science.Footnote 5 The Brown government attempted to limit success fees in defamation and privacy cases governed by ‘no-win-no-fee’ conditional fee agreements (CFAs).Footnote 6 After all three party manifestos at the 2010 general election promised libel reform, and the coalition agreement, the founding document of the new government, included a commitment to “review libel laws [sic] to protect freedom of speech”,Footnote 7 Lord Lester of Herne Hill introduced a private member's bill, which received a second reading after a debate in which many of their Lordships raised the issue of costs, using words such as “massive”, “excessive”, “enormous”, “devastating”, “crippling” and “out of all proportion”.Footnote 8 In the legal sphere, media and campaigning organisations intervened as third parties in the European Court of Human Rights case MGN Ltd v. UK Footnote 9 to argue that high costs in libel and privacy cases create a ‘chilling’ effect sufficient to establish a violation of article 10 of the European Convention on Human Rights.
The proposition that libel cases are too expensive is now so firmly established in the political debate that those concerned with libel reform usually take it for granted and move immediately to a discussion of what should be done about it. The Campaign for Libel Reform, for example, calls for a cap on claimants' lawyers' base fees in CFA cases, for the abolition of success fees paid by defendants in the case of a successful action brought under a CFA and for a lower-cost libel tribunal to decide most or all cases.Footnote 10 The Culture, Media and Sport select committee, having heard evidence from Sir Anthony Clark M.R. to the effect that universal cost-capping would involve pure guesswork about the reasonable future costs of a case and would in itself add to costs, rejected cost-capping as “crude”,Footnote 11 but instead favoured a Ministry of Justice-sponsored pilot of a system of cost budgeting, in which each side sets out a plan for its costs, which the court monitors.Footnote 12 The Libel Working Group referred to a “widespread perception that the costs of [libel] proceedings are prohibitive”Footnote 13 and discussed a number of procedural reforms to reduce costs.Footnote 14
This move to remedies is, however, far too fast. As this article sets out to show, the reality is that we know little about the costs of libel cases, and what we do know does not justify precipitate measures that would have the effect of reducing claimants' access to justice.
Evidence submitted to the Jackson Review
The one publicly available dataset of libel costs is a survey of 154 ‘publication’ cases – which is to say, libel and privacy cases – from 2008 contained in an appendix to the preliminary report of the Jackson Review of Civil Litigation Costs.Footnote 15 Only cases in which the defendants were media organisations were included in the survey, a restriction that arose out of the fact that the data were supplied by the Media Lawyers Association (MLA), a group, despite its name, not of law firms but of 22 major media organisations.Footnote 16 The Jackson Review produced its own analysis of the survey,Footnote 17 but, for reasons associated with its purpose of investigating the effects of CFAs on overall costs,Footnote 18 it concentrated on the 27 CFA cases identified in the survey and compared them with 34 selected non-CFA cases. The survey can also be used, however, as a source of information about libel costs in general.
At first sight, the survey seems to support the view that libel costs are very high. If one looks at the 139 cases that involved a defamation element (i.e. those that were not solely about privacy), the average costs incurred by the claimant amount to £65,604. As the Jackson Report itself noted, however, some of the cases were not at that stage complete, so the MLA had recorded not the costs awarded to the claimant but the costs claimed by the claimant. The latter are often much higher than the former. In addition, one case is clearly an outlier of extraordinary proportions, a case in which the costs of the two sides added together amounted to more than £3 million. If those cases are excluded, the average claimant's costs come down to £23,227, or to £18,053 if one further excludes the one case out of the remainder that had gone to trial.
Even £18,053 might seem a hefty bill, but there is a further important point to make. Even after the exclusion of the speculative figures and the outlier, the data are still heavily skewed,Footnote 19 so that a few very expensive cases exert a disproportionate influence on the value of the average expressed in terms of the mean. If one instead looks at the average in terms of the median, the value of the case exactly in the middle of the range, the figures are much very much lower: £5,023 for the set of all 139 cases, £5,000 if one takes out the speculative figures and the outlier, and again £5,000 if one further takes out the remaining case that went to trial.
If we turn to mean defendant's costs as recorded in the survey, a similar picture emerges. The mean for defendant costs across all 139 cases comes to £31,152, falling to £11,004 if one discounts the speculative figures and the outlier. Removing the remaining case that went to trial, however, has a bigger effect on the defendant side than on the claimant side and reduces the mean to £5,589.
Median values on the defendant side are not so simple. A large number of cases show defendant costs as zero. According to the Jackson Review, the explanation for this apparent oddity is that these cases were dealt with by the defendant's in-house lawyers. But there are so many zero-cost cases in the survey that the median also comes out as zero. Obviously, zero is not a realistic estimate of defendants' costs – in-house lawyers have to be paid for even if their cost is not recorded in the same way as that of external lawyers. If we exclude all the zero-cost cases, leaving 62 cases, we can obtain a better idea of defendants' costs. The mean value in these cases rises to £69,841 (more than mean for claimant costs in the survey as a whole, one should observe, even though claimant costs include CFA success fees in 27 cases), but falling to £27,729 when the speculative and outlier cases are removed, and to £14,258 if one ignores the remaining case that went to trial. With the zero-cost cases removed, we now have meaningful median values. They follow the same pattern as the claimant side: £6,310 for all the libel cases, £5,181 with the speculative and outlier cases removed, and £5,088 with the remaining trial case taken out.
If one were to ask about the total costs of a typical libel case, adding the costs of both sides together, an answer based on mean values for the two sides in all 139 cases comes out as nearly £97,000, and as around £135,000 if one uses the mean of the 62 non-zero cases for defendants' costs and the mean of all 139 cases for claimants' costs. But an answer based on median values, even with the speculative and outlier cases left in and including cases that went to trial, is just £7,497. If we take the median of the 62 cases as a meaningful value for defendant costs and add it to the median of the 139 cases for claimant costs, the total costs figure rises to £11,333.
The median values for claimant's costs and defendant's costs taken separately are similar – £5,023 for claimant's costs for all 139 cases, £6,310 for defendant's costs over the 62 cases for which we have plausible figures. That picture changes somewhat if we look at claimant and defendant costs simultaneously for each of the 62 cases. The median total costs for these 62 cases taken separately is rather high – £28,193, £21,884 of which comes from the claimant side. The median for claimant costs in the defendant zero-cost cases, however, is only £2,585. The difference is, as one might expect, even greater for the mean values (£61,945 versus £3,602). That is, claimants spend more when faced by defendants who hire outside lawyers than when faced by defendants who use in-house lawyers. Given that that some defendants probably deal with low value cases in-house but put high value or more difficult cases out to external lawyers (which is compatible with the fact that all the cases in which claimant costs came out at over £100,000 fall into the non-zero defendant cost category), no definite conclusions can be drawn, but the relationship is suggestive and certainly deserves further investigation.Footnote 20
Claimant costs and defendant costs are positively correlated. As one rises, so does the other. But, broadly, claimant costs start higher and increase at a rate approximately one and half times the rate at which the corresponding defendant costs increase.Footnote 21 If one takes out the 21 CFA cases included within the 62 cases, those which generate a success fee, so that we are looking at base fees plus disbursements for both sides, in the remaining 41 cases, although claimant costs remain higher than defendant costs (the mean for the former is £62,331, for the latter £45,802), they rise at a rate of only about one and one fifth of the rate at which defendant costs rise.Footnote 22 From this, although one can say that claimant costs are higher than defendant costs, the evidence comes nowhere near justifying the opinion of the MLA that there is “a huge disparity between the costs of claimants and those of defendants”.Footnote 23
The positive correlation between defendant costs and claimant costs is remarkably high.Footnote 24 Several possible explanations come to mind. One is that each side is influenced by the other side's activity and thus spends to keep up in a form of arms race. Another is that both sides are influenced by the amount at stake in the case. There is a further correlation between the damages awarded and the parties' costs, albeit not a strong one.Footnote 25 The data do not, however, clearly point to one explanation as better than the other, and it is entirely possible that both are operating at the same time.
Public discussion, including judicial discussion, of the correlation between claimant and defendant costs tends to assume that, especially in CFA cases, claimants' lawyers are engaging in the tactic of driving up costs to induce defendants to settle regardless of the merits of the case.Footnote 26 Given the complex burden of proof rules in defamation, the assumption that it is always the claimant who is taking the initiative seems rash, and the data are compatible with causation moving the other way, from defendant to claimant (as well as from a third factor such as the amount at stake). The possibility that claimants spend less when faced with a defendant represented by in-house lawyers would be an example of defendant behaviour influencing claimant behaviour rather than the other way round.
What conclusions can we draw from the MLA data? The degree to which the data are skewed makes conventional statistical analyses somewhat unreliable and we have the results of only one survey based on one year (albeit one produced by the media themselves and so not likely to be biased in favour of low values). But one can say with reasonable confidence that the total costs of a randomly selected libel action are more likely to be in the region of £10,000s than of £100,000s or of £1 millions. We can say that claimant costs are higher, but not vastly higher, than defendant costs. And we can say that there are a small number of cases that generate costs very much higher – even 100 times higher – than those in a typical case. Whether that equates to a situation in which libel costs are ‘too high’ is a separate question, to which we shall turn shortly.
Before moving on to that question, however, we should deal with one other study of libel costs, a study that has attracted much attention from libel reform campaigners, the so-called “Oxford University Study”.
The “Oxford University Study”
One of the most frequently cited figures about libel costs is that costs in England are 140 times higher than the European average. The figure appears both in political debateFootnote 27 and in argument before the courts.Footnote 28 Another is that English libel cases are four times more expensive than the next most costly jurisdiction in Europe, which itself is ten times more expensive than the third most expensive jurisdiction. Footnote 29 The source of these claims is a study compiled by Troels Larson and Danilo Leonardi, then of the Centre for Socio-legal Studies in Oxford. Following the somewhat odd journalistic convention that entire higher education institutions are held responsible for every piece of academic research that goes on within them, the study is known in political circles as the “Oxford University Study”.Footnote 30
These figures are, however, absurd. Even using the highest figure for average total costs that one can plausibly generate from the MLA survey, namely using the mean of all 139 cases for claimant costs and the mean of the 62 non-zero cases for defendant costs, the ‘140 times’ claim would imply that the European average for total costs in a libel action was under £1,000. Using the more realistic median figures, whether drawn from the same mix of the 139 cases and the 62 cases or just from the 62 cases, the implication would be that the European average for total costs in libel cases was somewhere between £79 and £195. This is highly implausible. It would mean, for example, that, even using the near £1,000 figure, average total costs of both sides in a libel action in the rest of Europe come out at less than half of the claimant's base fees in the average road traffic case in England,Footnote 31 even though such cases are very numerous and dealing with them benefits from considerable economies of scale. Using the lower figures, average total costs in the rest of Europe would buy about one hour of an English personal injury defence lawyer's time at 2007 prices.Footnote 32
Another point is that if libel were so cheap in the rest of Europe, other things being equal, one would expect to see extensive forum shopping away from England, as claimants made heroic efforts to establish jurisdiction in other European countries. Why a claimant in an internet case should choose England in which to pursue a case rather than, for example, Germany, where their costs would be limited to a few hundred euros, would become a very great mystery. England might have other advantages for claimants, in terms, for example, of substantive law, but cost advantages of such an enormous magnitude would surely induce some claimants to take the required risks. A claimant in a non-CFA case (the most likely scenario on the Jackson figures) faced with a two-thirds chance of success using English law might well rationally prefer a fifty-fifty chance, or an even lower chance, in a jurisdiction where the costs are typically a hundred and fortieth of English costs. No such pattern has been reported.
Even a cursory glance at the study quickly reveals the problem. The study's method is extremely weak and its conclusions unwarranted. The estimates for costs in libel proceedings in England were obtained by asking two lawyers, both of whom worked for the same top-of-the-range defendant-side firm, to put figures on costs for two anonymised cases. Both cases, it seems, were ones in which the lawyers had in fact acted for the defendant.Footnote 33 The lawyers came out with total costs of more than £550,000 in the first case and “up to” nearly £4 million in the second case. The same facts were then put to lawyers from various other countries, who were asked to estimate their total costs within those other systems.
The objections to this method are plain. The English lawyers and the foreign lawyers are not in equivalent positions. The English lawyers know the real cases. For the foreign lawyers they are purely hypothetical. The English lawyers, in consequence, did not treat the cases as average or typical cases, but as they really were, with all their complications. For example, the English report reveals that its estimate of costs in one of the cases assumes that there would have to be “a number of complex applications before trial”, presumably because this is what happened in the real case.Footnote 34 No equivalent assumptions appear to be made in the other reports. Moreover, the English examples proceed on the basis that the firms involved were the actual (evidently very expensive) ones, whereas the foreign lawyers were likely to interpret what they were asked to do as to give typical costs, not costs using equivalently expensive firms in their systems.
In addition, the foreign lawyers seem to have had great difficulty translating the idea of “costs” and “liability for costs” into their systems. Do “costs” include lawyers' fees, for example? And do they include process costs that cannot be recovered from the other side in the event of victory? In the French report, defendant “costs” in scenario one, which the English lawyers reported as £145,000, are given as “€0”,Footnote 35 even though the French report just previously gives defendant lawyers' fees in that scenario as €15,000.Footnote 36 Evidently, as we can see from the answer to the question about costs in scenario two, which mentions costs for “witness accommodation”, the French respondents interpreted defendants' “costs” to mean principally what English lawyers might think of as “disbursements”. In the next answer in the French report,Footnote 37 therefore, the respondents fail to include the defendant's own lawyers' fees in the defendant's cost “liability”.
Moreover, the French report points out that the scenario two case would be treated as a criminal matter in France (since, in France, as in many other jurisdictions, defamation is principally a crime and only secondarily a matter of civil liability). Nowhere in the pro forma, however, is there any space for reporting the costs (recoverable or otherwise) incurred by the state.Footnote 38
Some reports lack any specific information on actual costs and give only the legal rules concerning costs. In Germany, for example, as is clear from the answer on hourly rates,Footnote 39 German clients and lawyers may negotiate fees of any level as long as they are above a statutory minimum and are reasonable. But all the answers in the German report are given in terms of statutory minimum fees, not real fees, a response presumably made natural to German respondents because cost-shifting of lawyers' fees in Germany is restricted to the value of the statutory minimum fees.Footnote 40 That does not mean, however, that in the real world, German newspapers would invariably pay their own lawyers the statutory minimum fee, precisely €750, to defend a case that even in France would cost defendants €50,000.
Worst of all, however, is the way the national reports are summarised. In the survey's summary bar charts the German statutory minimum fees are treated as if they were the actual fees in real cases.Footnote 41 The treatment of the French report is, if anything, even odder. The French report gives a minimum of total lawyers fees (excluding state prosecution costs) as €20,000 in scenario one and €65,000 in scenario two, plus court costs and disbursements of €500 and €2000 respectively. But the summary bar charts give figures of £11,500 and £38,000, an implied exchange rate of £1=€1.74 in scenario one and £1=€1.71 in scenario two. The reality in December 2008, when the report was published, was that the pound and the euro were near parity.
The Larson and Leonardi study is, one is forced to conclude, nowhere near the quality required for policy purposes. Despite constant repetition of the “140 times” claim, in truth, no reliable figures seem yet to exist to facilitate comparison of libel costs between England and other jurisdictions.Footnote 42 There is some evidence that lawyers' fees in civil liability cases generally, not just in defamation, are higher in common law jurisdictions than in civil law jurisdictions, but there are offsetting differences in court fees, in the work lawyers do and in the costs absorbed by the state.Footnote 43 There might be grounds for asking whether other jurisdictions strike a better balance between cost and quality of justice across the whole of civil proceedings, but since British lawyers are not even the highest legal earners in Europe,Footnote 44 it would be surprising if savings measured in terms of orders of magnitude are available. Above all, the evidence does not establish that there is any specific problem about libel actions.
How much is too much?
Assuming that that the MLA survey results might be expected broadly to be reproduced in any larger scale, multi-year study, what are we to make of the current situation? Is the median libel case (£11,000−£27,000 of total costs) too expensive? How are we to judge? Such sums are considerable for an individual on median earnings (about £21,000 p.a. in 2008–09Footnote 45) but are trivial for media corporations of the size of News International (whose turnover in 2009–10 was £617.9 millionFootnote 46).
The Costs to Damages Ratio
One theme of the recent debate has been to compare total costs with damages. The MLA pointed out to the Jackson Review, for example, that total costs across the 154 cases it had identified in its survey were three times the total damages awarded.Footnote 47 Perhaps it would be more accurate to exclude the privacy-only cases, the zero defendant cost cases, the cases in which there was a global settlement including both costs and damages and the cases in which there were no damages at all. On that basis, using the resulting subset of 50 cases, total costs were four and half times total damages.
But is the ratio of costs to damages a useful measure of social utility? The problem is that for it to count as an accurate measure, one has to assume (1) that damages levels accurately reflect the harm caused by libel and (2) that the law has no deterrent effect whatsoever, so that one cannot count as an additional benefit of the law any harm avoided as a result of the law's existence. There is reason to doubt not just one of those assumptions, but both of them. The difficulty with the first assumption is that damages levels are artefacts of the law itself. Ever since Rantzen v. MGN Footnote 48 and John v. MGN,Footnote 49 the appellate courts have been exerting downward pressure on libel damages and attempting to rein in the instincts of juries to make large awards. How are we to decide whether the judges have identified the true social value of the harm? Indeed, one can turn the argument on its head and point out that since the parties are clearly willing to spend far more either in defending or in attacking reputations than the monetary amounts at stake, they must be placing a higher value on reputation than that implied by the level of damages. As for the second assumption, since media organisations themselves repeatedly complain about the “chilling effect” of libel law, they are hardly in a position to assert that the law has no deterrent effect.
Access to Justice and the Importance of CFAs
Another way of approaching the question of how much is too much is to consider the matter from the angle of access to justice. On the assumption that large media corporations such as News International can look after themselves, there are two sets of potential litigants to consider: individual claimants and arguably vulnerable defendants such as allegedly struggling local newspapers. In both instances, the issue is whether the likely costs are sufficient to induce a party with a strong case, either as a claimant or as a defendant, to abandon it because of the risk of a mistaken adverse judgment. With regard to individual claimants, it is quite possible that median costs in the range of £11,000 to £27,000 would deter all but the well-off. Legal aid has never been available for defamation actions,Footnote 50 which was presumably one of the reasons that defamation developed a reputation as a rich plaintiff's tort. The position of individual claimants should nevertheless have been improved by the introduction of the CFA system, which allows claimants to limit their risk to the cost of disbursements by their own lawyer and the cost of after-the-event-insurance (ATE) premiums. Sometimes they do not face even the risk of that cost, since ATE policies often include cover for disbursements and some lawyers are willing to absorb ATE premiums as a cost of their own business. The principal barrier to access to justice under the CFA system is the willingness of a lawyer to take the claimant's case in the first place, which in turn depends on the balance between the lawyer's estimate of the chances of success and the uplifted fees the lawyer is permitted by the CFA regulations to recover from the losing defendant.
The MLA survey data illustrate the effects of CFAs. Of the 139 libel cases, claimants failed to win any compensation in 47 cases. In the rest they either received damages or a global sum rolling damages and costs together. The claimant used a CFA in 27 of the 139 cases. In those 27 cases, the claimant failed to receive compensation in only 3 cases. In other words, using a CFA increased the chances of a claimant receiving compensation from 66.2% (92 wins out of 139 cases) to 88.9% (24 wins out of 27 CFA cases). Admittedly, these sample numbers are low, and it is not clear what explains the figures. Do claimants' lawyers put more effort into CFA cases, or is it that, as anti-libel campaigners sometimes suggest, defendants are intimidated by claimants funded by CFAs because they know that CFA claimants face low downside risks, or is it merely that lawyers choose to allocate their better cases to CFAs because they are able to charge higher fees? We return to these topics below.
In passing, one should mention an odd, and potentially misleading, claim by the Campaign for Libel Reform about what the MLA survey shows about the chances of a claimant winning a libel case. The Campaign said in its briefing to members of the House of Lords for the debate on the second reading of Lord Lester's Bill that “out of 158 [sic] cases from 2008 identified in Justice Jackson's [sic] review of civil litigation costs, none was won by the defendant”.Footnote 51 Whether that is true turns on the precise meaning of “won”. It is true that of the 139 libel cases in the survey, only two went to trial, and both of those trials were won by the claimant. But the fact that claimants received no damages in 33.8% of cases is probably a better measure of defendant success rates than the 0–2 score for trials apparently being used by the Campaign.
On the defendant side, the claim is often made that local newspapers have stopped defending libel actions because they cannot afford to fight them.Footnote 52 The MLA data, however, throws some doubt on that claim. Plainly, local newspapers might not survive a case with costs in the millions, or even in the hundred thousands, but the data indicate that the vast majority of cases lead to costs in the tens of thousands. The defendants in the survey's median libel case faced a total final bill – their costs plus the claimant's costs plus the damages – of £14,458. That is an annoying amount for a local newspaper to have to pay (enough to be a deterrent, one would hope) but it is hardly “devastating” or “crippling”. Local and regional newspapers made operating profits of £495 million in 2008, the same year as the MLA survey.Footnote 53 It would take 34,238 median value libel cases in a year to wipe out that profit. Even using mean values the figures would be £123,416 and 4,011 cases.Footnote 54 The idea that local newspapers are being overwhelmed by the cost of libel actions seems deeply implausible. Moreover, the risk to each individual publication can be managed through liability insurance, which is widely available.Footnote 55 If it is indeed the case that local newspapers defend very few cases, another possible explanation might be that reductions in the resourcing of journalism have reduced the quality of local newspapers so much that they find themselves in situations far more often than previously where their position is legally indefensible.Footnote 56
It is also worth pointing out that defendants are just as able as claimants to proceed on the basis of a CFA. CFAs are not the same as contingent fees. They do not depend on the existence of a pot of damages out of which only the claimant's lawyers can be paid. There is no obvious reason why defendants should not take advantage of them.
Problems of CFAs
Anti-libel campaigners complain that, since rich claimants use CFAs as well as poor ones, access to them should be limited. Means-testing is a possible, though complex, response to that point, but many of the other possible restrictions of CFAs would hit poor claimants as well as rich ones. It is even possible that they would affect poor claimants disproportionately. The central objection to the former government's proposal to cut the maximum success fee in libel cases from 100% to 10% was that its effect would be that claimant's lawyers would weed out more and more cases as a bad bet.Footnote 57 The weeding process would not take into account claimants' means, only the chances of success. On the other hand, any cases lawyers accepted instead on the traditional fee basis would be likely to come from better-off claimants. The result is that, whatever the mix of claimants under the present CFA arrangements, that mix would be more skewed to the wealthy if success fees were reduced.
There are also objections to the weeding process itself.Footnote 58 If the allowable success fee is, for example, 100%, lawyers will make at least as much money from success fees as they would have made from traditional fees as long as the chance of success in the cases they take is at least 50%. If they weed out cases with a chance of success lower than 50%, but take those with more than a 50% chance, they will make more than they would have made in traditional fees. The complaint is that success fees are meant to fund cases with a lower chance of success, but in practice such cases are weeded out and lawyers appropriate the net proceeds to themselves.Footnote 59 The problem with the complaint, however, is that one person's scandalous ‘cherry-picking’ is another person's deliberate act of policy. The level of the success fee – whether controlled ex ante by regulation or ex post by a costs court – determines the point at which lawyers are incentivised to weed (or ‘cherry-pick’). If the authorities want cases with a less than 50% chance of success to be taken forward by CFAs, they could set the success uplift at greater than 100%. Since they have chosen not to do so, the implication is that public policy intends to exclude such cases from CFA funding.
Another alleged weakness with CFAs is that there is no requirement that the claimant take out ATE insurance. Impecunious claimants, unable to pay the defendant's costs in the event of losing the case, might nevertheless find lawyers willing to run up costs early on in the hope that the defendant, observing that the claimant is judgment-proof, decides that early settlement is a better prospect than winning expensively.Footnote 60 In some eyes, this is a form of blackmail.Footnote 61 The problem, however, with this game of legal ‘chicken’, as with the ‘nuisance suit’, its counterpart under the American costs rule,Footnote 62 is that the claimant's lawyer also faces unrecoverable costs. If the defendant refuses to buckle, the claimant's lawyer will have wasted time that could have been spent on remunerated work, or on CFA cases with better chances of success. Since the claimant is impecunious, the claimant's lawyer will not even be able to recover disbursements. It is not therefore surprising that evidence is lacking that the problem is widespread. When it has allegedly appeared it has been dealt with by a cost-capping order.Footnote 63 Perhaps a more general solution would be to restrict CFAs to claimants who take out ATE insurance or to provide that where there is no ATE insurance, defendant costs become recoverable from claimants' lawyers, but whether the problem is serious enough to justify any intervention is far from clear.
Other problems with the CFA system are less easily countered or solved, however. The clearest is that CFAs result in no market pressure on claimant fees.Footnote 64 No-win-no-fee claimants never pay the fees they agree with their lawyers. The only people who pay are losing defendants, who do not agree the fees. The consequence is that the only rational strategy for claimants' lawyers under no-win-no-fee is always to set fees at the maximum they think they might induce a costs judge to endorse or any applicable statutory maximum allows. The result is that CFAs put more pressure on the system for regulating fees than it was designed for, and a more effective system might be required.Footnote 65
As Jackson L.J. pointed out to the Culture, Media and Sport select committee of the House of Commons,Footnote 66 the point about market pressure applies to all civil cases, not just to defamation cases. If there is a problem peculiar to libel CFAs, it might be the cost of ATE premiums. The Jackson Review's final report refers to ATE premiums in libel cases of £65,000 for £100,000 of cover, which seems very high in comparison with other types of case.Footnote 67 It is also striking that although many lawyers seem to have taken for granted a figure of about 50% for the proportion of libel cases taken on a CFA basis, the MLA survey could only identify 19%. It is not clear why any of this should be the case, but if it represents the normal situation facing libel claimants, it would suggest that there is still an access to justice problem in libel despite the existence of the CFA system, a problem that relates not to claimant costs, but to defendant costs.
One possible explanation is the extent to which defendant libel costs are skewed.Footnote 68 A very small number of cases generate very high costs, costs that have to be met from the pool of ATE insurance premiums. The consequence might be that insurers set premiums that strike the vast majority of their clients as too high for the risk as the client sees it. Unfortunately, we do not have defendant cost data for other types of case that might allow us to test this idea. One point against it is that claimant costs for other types of case are also heavily skewed,Footnote 69 but we do not know whether the effect carries over to defendant costs, and so we cannot compare degrees of skewedness in different types of case with what we can find out about CFA use.
Human Rights Limits
Even if access-to-justice problems persist in libel actions despite the availability of CFAs, it would seem a perverse response to those problems to restrict access to CFAs, since the likely effect would be to restrict access to justice even further. But that is precisely what the European Court of Human Rights seems to have decided in MGN Ltd v. UK.Footnote 70MGN concerns privacy rather than libel, but its reasoning covers both areas.
The facts are familiar to students of privacy law. The Daily Mirror surreptitiously took photographs of the celebrity model Naomi Campbell leaving a Narcotics Anonymous meeting. The House of Lords held, by a majority, that publication of the photographs and of details of Ms Campbell's treatment for drug dependency constituted a breach of confidence and violated her right to private life.Footnote 71 The defendants' human rights challenge to the substantive decision failed before the Strasbourg court, but they also argued that a violation of article 10 arose from the success fees awarded to the claimant in respect of the proceedings before the House of Lords, both on appeal and in subsequent proceedings in which the defendants challenged the first set of success fees.Footnote 72
The newspaper pointed out that excessive damages in publication cases could amount to a violation of article 10,Footnote 73 as could the mere possibility of unpredictably high damages.Footnote 74 It followed, they argued, that the award of very high costs could also amount to such a violation. The newspaper, however, made no complaint about having to pay the claimant's base fees. It argued only that the award of success fees was contrary to article 10, on the grounds that success fees pursued no legitimate aim and lacked proportionality. They submitted that success fees were by definition disproportionate, because they were greater than the claimant's actual costs in vindicating her rights, that total claimant costs (agreed at £500,000) were out of proportion to the damages paid (£3,500), that the lack of any incentive on the part of claimants in a CFA to keep costs down meant that the success fees (but, oddly, not the base fees) were excessive, that the Jackson Report and the consultations and debate surrounding it in Parliament and in government demonstrated that controls on CFA costs were deficient, that the threat of blackmail suits had a ‘chilling effect’ on the media, that the possibility of ‘cherry-picking’ meant that CFAs did not help impecunious claimants gain access to justice but simply transferred costs from legal aid to defendants, and that allowing wealthy claimants such as Naomi Campbell to use CFAs, when means-testing was in operation in other parts of the legal system, demonstrated that CFAs did not serve the legitimate end of providing access to justice for the impecunious.
The Court rejected the newspaper's contention that CFAs served no legitimate aim. The relevant legitimate aim was to provide “the widest public access to legal services for civil litigation funded by the private sector and thus the protection of the rights of others within the meaning of Article 10 §2 of the Convention.”Footnote 75 But the court accepted that the award of success fees did not constitute a proportionate means of achieving that aim, and so were not “necessary in a democratic society”. It did not, however, follow the newspaper's argument that any award of costs above the base costs is “disproportionate” or that the costs-to-damages ratio is relevant. Instead it proceeds from the Jackson criticisms of the existing scheme – in particular, the lack of control on who might use a CFA together with the scheme's “lack of focus”, the lack of any incentive to control costs, the threat of “blackmail” cases and the “cherry-picking” argument. The court notes that the UK Ministry of Justice seems to agree with most of these criticisms, but that so far no major changes have occurred. It then announces:
[T]he depth and nature of the flaws in the system, highlighted in convincing detail by the public consultation process, and accepted in important respects by the Ministry of Justice, are such that the Court can conclude that the impugned scheme exceeded even the broad margin of appreciation to be accorded to the State in respect of general measures pursuing social and economic interests.Footnote 76
A number of questions might be raised about this decision. Is it legitimate for a supranational court to intervene in such a detailed way in a policy decision? Is it prudent to use the fact that a policy is being reviewed as a ground for saying that it violates human rights, given that governments might in consequence refuse openly to review problematic policies? Footnote 77 Those debates, however, though important, would take us far from the topic of the current article.
Of more immediate concern is the question of what MGN Ltd v. UK says about what might count as costs that are too high. It provides no quantitative guidance. The court objects to success fees, not to any particular total sum awarded in costs, even though cases exist (for example in the MLA survey) in which the claimant's base fees exceeded the sum of base and success fees agreed in MGN. Another puzzle is that, despite the importance attached to the defendant's means where an article 10 violation arose out of the size of the damages awarded,Footnote 78 the court in MGN makes no reference to the newspaper's available resources, or to its ability to manage risks using liability insurance.Footnote 79
The court's objections seem more qualitative than quantitative (although how such objections could rationally relate to “chilling” effects remains somewhat obscure). Its reference to the lack of entry qualifications for CFAs and to “lack of focus” in the regime suggest that a scheme aimed more precisely at widening public access to the courts (for example by excluding through means testing litigants who would enjoy access in any event) might pass muster, even if the costs generated were no different. The lack of incentives for claimants to restrict costs might be met by stronger direct regulation, and that might reduce costs, but it would make little sense to regulate success fees more rigorously without similarly regulating base fees, since success fees are expressed as percentage of base fees.Footnote 80
The court's references, however, to the “blackmail” and “cherry-picking” points are more troublesome. It is far from clear that either is a real problem. As for solutions, it might be possible to prevent instances of “blackmail” within a CFA regime in the ways suggested above (no CFA without ATE or claimant lawyer liability). Neither would reduce costs per case but both might reduce the number of cases brought. “Cherry-picking”, however, is more difficult to deal with because it is arguably an essential feature of the scheme, not a bug in it. Indeed, if the court is saying that only policies that fund weak cases brought by poor claimants are proportionate, states will have great difficulty finding an acceptable policy. Only two possibilities spring immediately to mind: exceptionally generous legal aid and compulsory pro bono work. The former seems unlikely to find favour in any currently foreseeable economic conditions,Footnote 81 and the latter has its own human rights complications.Footnote 82
Perhaps the most disturbing aspect of MGN Ltd. v. UK's treatment of “blackmail” and “cherry-picking” is that it threatens Jackson L.J.'s suggested replacement for CFA. No-win-no-fee arrangements would continue to be possible but defendants would no longer pay success fees and, crucially, claimants would not have to pay defendants' costs in any circumstances, even when defendants win. Defendants who pay their own costs are inevitable targets for ‘blackmail’ (or “nuisance”) litigation.Footnote 83 The problem would persist, if not worsen. Moreover, lawyers would continue to “cherry-pick” in the sense condemned by the court. They would offer no-win-no-fee deals only to claimants with a worthwhile chance of winning, which would mean that defendants would face paying their own costs, perhaps high enough to give rise to a “chilling” effect, in a subset of cases chosen by claimants' lawyers on the sole basis of their chances of victory.
Comparison with Costs in Other Types of Case
Another way of assessing whether libel costs are too high might be to compare them with costs in other types of case. Here we can again turn to the evidence collected by the Jackson Review. According to a summary (unfortunately rather a bare one) of a survey of 15,000 personal injury cases,Footnote 84 average claimant costs in personal injury cases are around £3,000 to £4,000. Another Jackson Review submission suggests that one should approximately double those figures if the case goes to trial.Footnote 85 Defendant costs are more obscure, but assuming that they are, at most, about the same as claimants' lawyers' base fees, a range of £2,000 to £2,300 seems about right. In contrast, the average (mean) claimants' costs in a small collection of Chancery cases presented to the Jackson Review came out at over £80,000 (median: £58,000).Footnote 86 Jackson also organised users of the Commercial Court to collect evidence of the costs of cases in that court.Footnote 87 Average claimant costs in those cases seem to be over £300,000 (median: £90,000), sums perhaps surprisingly described by Jackson L.J. as “proportionate”, by which he meant that claimant costs were often less than the amounts at issue.Footnote 88 Another collection of cases, albeit heavily weighted down with disputes that had gone to judgment, came from the Technology and Construction Court.Footnote 89 Average claimant costs in those cases were as much as £448,932 (median: £63,784). Overall, Jackson identified 143 Chancery, commercial, technology and construction cases in which average (mean) costs recovered by claimants were £360,834 (median: £87,000). Information was not gathered about defendant costs, but there is no reason to believe that, apart from any success fees in CFA cases, they would differ greatly from claimant costs, and so it would be reasonable, in the absence of further information, to estimate total costs at just short of twice claimant costs to arrive at an overall average of roughly £700,000 (median: £170,000). If one takes out the cases where there is evidence that they involved a full trial or an appeal, the average falls, but only to £364,711 (median: £77,500).
If we ask whether costs in libel cases are too high in comparison with costs in these other types of case, it becomes apparent that the question is a complex one of relative values, rather than an accounting exercise. Libel cases cost more than personal injury cases but, it seems from the very limited evidence we have, much less than disputes about property and commercial contracts. Is that justified? Some might argue that personal safety is inherently more important than reputation and that, consequently, no more should be spent on libel than on personal injury. But that argument ignores the availability of economies of scale in personal injury litigation, and, more importantly, would apply a fortiori to property rights and commercial disputes. Why should commercial disputes use up between 5 times (mean values) and 15 times (median values) more resources than libel cases? One might counter that that the vast sums spent on commercial disputes are justified by the fact that the parties are willing to spend that much, but precisely the same argument would apply to libel cases.
There is, it seems, no easy route from a comparison of libel costs with the costs of other types of case to a convincing argument that libel costs are ‘too high’. Such a comparison does not answer the question ‘too high for what?’
The Very High Cost Cases
But, finally, what about very high cost cases, the ones that greatly skew the data? Surely we can say that their costs are too high? Four cases in the MLA survey produced total costs of over £1 million, two of which were around £1.5 million, one over £2.5 million and one, £3.14 million. The next highest total after these cases was £458,720. In two of the £1 million-plus cases, the figures are speculative, because they were based on what the claimant had asked for rather than what the claimant had been awarded, and in one case the costs are estimated, but that still leaves one case with costs over £1 million, a case which, the notes to the survey tell us, was subject to a costs cap.
The difficulty with saying that costs even in these cases were excessive is that we know very little about them. Two of them followed a trial, which obviously is very expensive way of resolving a dispute, but the other two were settled before trial. Three were CFA funded, but one was not. The damages awarded were substantial, ranging from £85,000 to £300,000, but their costs to damages ratios were much worse than average – on average in these cases, total costs were more than 11 times damages. Other than those scraps, the survey tells us nothing. It might be possible, using a certain amount of detective work, to track down cases that seem to fit these sketchy profiles, but the results would of necessity be speculative.
Is there nevertheless a case for a long-stop limit on costs in libel cases, on the ground that whatever their details, and whatever the human rights position, total costs of this order are a waste of society's resources? The problem is that, even if we lay aside the justification that the parties themselves consider the costs to be worth it, or at least worth more at the time than the consequences of conceding, there might be social benefits generated by such litigation. Very high cost cases might, for example, be test cases about extremely important points of law. Two of the four survey cases were settled, and so seem not to fall into that category, although it is possible that they were settled after legally important pre-trial hearings or even after an appeal.
It is admittedly not easy to determine the social value of resolving a point of law, but one should be wary of declaring, especially in advance of the evidence, that no point in libel law is worth more than a specific sum to decide. To the extent that libel engages fundamental rights, such as freedom of expression and the right to a private life, a rigid cap approach seems difficult to defend.
Another justification for very high costs might be that the damages do not fully capture the importance for the claimant of the harm caused or the issues at stake. Examples of the former might be accusations of child abduction and murder, as in the McCann litigation,Footnote 90 or of Al Qaeda-related terrorism, as in King v. Telegraph Group.Footnote 91 Examples of the latter might include the Holocaust denial cases, such as Irving v. Penguin Books.Footnote 92 Again, to declare that no such case can ever be worth more than a particular amount in costs seems over-hasty and to risk very great injustice.
Conclusion
What can we say about libel costs in the light of the evidence, such as it is? We can probably say that libel costs are heavily skewed – with a small number of very expensive cases dominating both the calculation of the average cost of cases and the headlines. And we can also be reasonably confident that the vast majority of libel cases cost relatively little. They probably cost more than personal injury cases but less than commercial disputes. Beyond that, we can say very little.
We can perhaps be more definite about what we do not know. We do not know how costs in libel cases in England compare to those in other countries. We do not know whether there is still a problem of access to justice, despite the creation of CFAs. We even know little about the proportion of libel cases funded by CFAs or which kinds of case turn out to be the very expensive ones.
But above all, we do not know enough to justify a moral panic about libel costs. There is no evidence that libel costs “regularly reach hundreds of thousands of pounds” or that they are “out of control”. Perhaps one thing we can say with some degree of certainty is that the fervent confidence of anti-libel campaigners should be replaced by a fervent search for reliable evidence.
