Firm A goes up against firm B in a billion-dollar international arbitration. Firm A wins. But its client receives far less money than it was expecting. In parallel, B is pitching for the job of defending an ICSID claim against C. This time, B succeeds.
Who has the bragging rights, if any, between A, B and C?
If the context were sports, fans of A would surely gloat over B and fans of B the same over C. In international arbitration, things aren't so simple. A real-life arbitration helps to illustrate.
Once upon a time, the leading international arbitration firm of its day - Coudert Brothers - represented an Asian conglomerate in an apparently minor arbitrated dispute with a Middle Eastern agent. Except - and here's the rub - the agent wanted vested rights in most of the Asian firm's businesses, projecting the effect of those rights forwards ad infinitum.
So, although the dispute crystallised around a single contract claim, for $50 million, the stakes were plainly much higher. If the ICC tribunal were to accept the agent's position in one contract it would ripple across the conglomerate's balance sheet forever. To make matters worse, most of its big corporate loans contained acceleration clauses in them and could need early repayment in that event. Because of those, any legal award of more than $5 million had the potential to affect its very financial foundation.
Against that backdrop, the corporation in peril put out an offer of $16 million in full settlement: it was rejected. It had no choice but to go to trial.
As things turned out - after a "pretty bloody fight" in one person's telling, "involving my first experience with forged documents" - the arbitrators awarded the claimant $2 million and declared the contract terminated.
"A silly summary of the case," says Jan Paulsson, who was on the team for Coudert, "would have been that we lost a little case." In reality, he notes, the entire team - plus spouses - were treated to an all-expenses paid trip to the Olympic Games (the client being an official sponsor) as guests of honour.
So, to return to the A and B scenario of the opening, A obtained some form of positive result, but, really, B won.
Similarly, when it comes to winning new business - the B versus C scenario - where B appears to 'win', context matters too. Sure, some clients appoint on merit, but many - particularly governments - tender according to price. Did B low-ball its offer? Possibly. As for firms who brag about the number of 'billion dollar cases that they're working on, well, it's been notable over the years how few of those, especially those filed in the public forum of ICSID, have produced anything like that amount in damages. Some of the largest are in the region of $150 million. One, so-called, billion-dollar claim proved to be so hollow that it actually flipped into negative territory. In that case - brought against Turkey (defended by Freshfields Bruckhaus Deringer) - the arbitrators cancelled the claim for lack of substance and hit the claimant with a punishing legal costs to deter others. The Polish claimant had to pay the state's legal costs, worth $4 million.
Despite all these nuances, at least one ranking of arbitration law firms uses the value of awards and the size of the claims that they're working on. The information it unearths is of course fascinating about the cases underway. But it's questionable how much else it reveals, about the relative merit of firms A, B and C.
If size-of-award or value-of-claim don't help, is there another way, that does?
A few years ago, Global Arbitration Review journalists and the magazine's editorial board gave thought to precisely that question. A better option they decided was to count merits hearings.
* * *
The merits hearing is when the rubber truly meets the road, from the perspective of the lawyers on a case. After months, perhaps years of skirmishing, the two opposing teams of lawyers decamp to wherever it's been agreed the neutral territory is and begin their final preparations. For the lead advocates that might mean isolating yourself with a cold towel on your head for several weeks as you absorb the entire case and mentally rehearse. You may run mock direct and cross-examinations. For younger lawyers it probably means logistics, checking all of the materials, testing the technology works, down to ensuring the hotel can provide catering at the hours it will be needed, which can be quite late at night.
From the research perspective, the merits hearing has a big advantage as a metric for vetting firms. It's predictable. Just as fissile material emits radiation, so an active arbitration team can't help but produce merits hearings. "If the issue is to measure the size and vitality of a practice then there's no better measure than the number of final hearings," said a member of the editorial board when asked to recall the debates that led to the GAR 30.
"That number instantly creates a picture of how busy a practice is and gives you a fairly good indication of its size without having to reference the number of self-proclaimed arbitration lawyers. It is also the metric least likely to be manipulated.
"Yes, there are cases where a tribunal may bifurcate the quantum phase - so you end up with two hearings - but this is marginal stuff. The only real drawback is that you lose track of cases where the matter settled shortly before the hearing - which isn't uncommon - or the tribunal declined jurisdiction in the first phase.
"Admittedly, no metric is perfect, but this approach is likely to be the next best."
* * *
With that philosophy in mind, the GAR 100 research team published the first survey that tracked firms' merits hearings three years ago. The first GAR 30 - which announced itself as "what the Fortune 50 is to the Fortune 500" - ranked international arbitration practices on the basis of the number of merits hearings that had taken place under particular arbitration rules, plus some other criteria (related to the experience and reputation of individual members).
For the second edition, the research team split that single table in half: producing one table that tracked merits hearings (the GAR 30) and a second that tracked what peer reviews research said about firms, plus experience.
You are now reading the fourth iteration of the GAR 30. The basic format has been retained but some additional columns have been added. The reasoning goes that, as when upping the resolution on a printer, the more lines of information are used, the crisper and clearer the image.
The firms are arranged according to a "score" that's a multiple of the performance in several columns. We don't include the final score these days as it's the order more than anything that matters.
To some extent, the results of the GAR 30 speak for themselves. A glance at the rankings will give you a snakes-and-ladders impression of the market over the last year - who's up, who's down and by how much - and that, more or less, is what the table is meant to offer. But to draw intelligent conclusions from these movements, it helps to have an understanding of what the GAR 30 formula actually measures. The answer of course, is merits hearings - but these days there's a bit more to it than that. A comparison of Debevoise & Plimpton and Bae Kim & Lee will helps to explain. It so happens that these two firms were co-counsel on one of the largest ICC matters in recent years (they won). More importantly - for present purposes - they had an equivalent score in the column for merits hearings, with eight apiece. Yet they are 25 apart. So what else is going into the equation?
The first differentiator is hearing size. The formula looks at, among other things, how many times the firm was on its feet in "bet the company" matters (see the column of same name in the table). More accurately, that column should be called "super-high-value cases" - each year we set an arbitrary value that corresponds to a sum-that-would-be-very-hard-to-absorb. And if the size of a claim exceeds it, the case is counted in the bet-the-company column. This year the value was $900million. Whether the client's existence is really in peril doesn't matter. (However, the sum that's more than $900 million does need to be a real sum - if, for example, arbitrators award less than the, say, $5 billion originally requested, and our researchers are aware of that, then they use the actual award as the value of the case, not the original claim.)
Second, the formula now tracks the total-value-at-stake during the research period in hearings (by adding the size of all of the claims that went to a hearing together.) This step flows from an observation that some firms have a steady diet of very valuable cases, but that in any given year, those might not trigger the >$900 million column.
So the more substantial the firm's work, on a regular basis, the higher in the table it will tend to appear. In addition, responding to feedback, the formula once again factors in how many treaty-law matters a firm is on. (This was an element of the very first edition of the GAR 30.)
Turning to our two guinea pigs, what does this mean? Debevoise & Plimpton performs strongly in two columns - bet-the-company hearings and total-value-at-stake - where its co-counsel does less well. It also has a clutch of treaty-law cases, whereas Bae Kim & Lee has none. Those differences explain the different final positions. Debevoise & Plimpton illustrates another aspect of the GAR 30 as a ranking. A firm can yo-yo around year on year, if its practice has a certain type of cross-section. In Debevoise & Plimpton's case, it works nearly exclusively on huge cases, with next to no filler or smaller items. As a result, its hearings pipeline can be described as "lumpy". Its ranking tends to depend on where those kinds of long, slow-burn matters are in their life cycle. A year ago, for example, when it was no less active but had been working on jurisdictional phases and interim measures, Debevoise was 21st in the ranking. The arbitration team would have billed the same number of annual hours. WilmerHale and Shearman & Sterling and a few other firms are in the same boat (although they have larger portfolios that smooth out the effect). A firm could also be at a disadvantage under the GAR 30 formula, if, for some reason, its work tends to settle more. So far nobody has shown any evidence that some practices have a higher settlement rate. Until then, the GAR 30 assumes settlement is a constant factor for all.
In other words, these days merits hearings alone aren't enough to put a firm at the top of the table. That said, if anyone wishes to re-rank the table that way - or according to the number of pending cases, or the value of current portfolio, or any of the other columns - they're more than welcome to. We've simply devised one formula by which to sort the data - reasoning that volume and value are both good characteristics in an international arbitration practice.
In the end, and despite these changes, the outstanding firm is again Freshfields Bruckhaus Deringer. With 39 merits hearings in the past two years, it can boast having appeared in front of more arbitrators than any other firm. In addition, as the figure in "total value at stake" shows, only Shearman & Sterling was more entrusted with client money.
What is the magic ingredient behind Freshfields' success? GAR has heard a few theories over the years. Some say it was in the right place at the right time (it was one of the first UK firms to establish a separate international arbitration group). Others cite strong internal practice management that ensured all arbitration cases were directed to this central group. But Freshfields insiders point to other factors. First, and most obviously, they say, they have a strong global network. Yes, some of their rivals are equally strong in one or two cities, but none at present can match Freshfields strength in so many of the key arbitral centres: Paris, London, and New York (not to mention its presence in Asia and the Middle East). In each location, the firm has more than one distinguished partner. How has it managed to build out to so many places? Although Freshfields associates are sometimes poached by rival firms, it's remarkable the degree to which key figures have stayed. The current crop of practice leaders have been with the firm for a decade or more. Furthermore, other firms have had (at times) equivalent global networks, but haven't managed to keep them in tact. A lot of credit for Freshfields' rise to this level of stability and dominance goes to Jan Paulsson, who himself chose to remain with the firm to shepherd through the next generation of talent. The group is also known among junior lawyers as a place where they will be given opportunities to pursue the vocational side of the discipline and to get away from the trenches of case work, at times. A year ago, an associate who's worked at a few international arbitration shops and is now very happily established within Freshfields, told a GAR reporter that at his current firm, you are encouraged to have a passion "for the greater development of international arbitration." Asked where that attitude comes from, he said "It all starts with Jan."
Below Freshfields, this year's 30 has a slightly different look. Indeed, the order is possibly closer than it's ever been to the list that an arbitration aficionado might write, if asked to jot the current leaders in the field on the back of an envelope. Shearman & Sterling, White & Case, WilmerHale and Debevoise & Plimpton are now the chasing pack, with Allen & Overy and King & Spalding not far behind. (Last year it was Shearman & Sterling, but then Hogan Lovells, Clifford Chance, and then White & Case).
The shuffling of positions, as always, partly reflects that it can take years for an arbitration to run its course and reach the merits phase. A practice could have started to be super-busy two years ago and, in the universe of the GAR 30, the light from that change would only be reaching the photographic plate now.
With that in mind, it's interesting to note that two of the firms that have expanded most visibly in this field by recruiting lateral partners and associates - Allen & Overy and King & Spalding - are now hovering at 7 and 8 in the ranking. Their figures in all columns are only likely to get bigger as the effect of their enlargement feeds through (although our researchers do their best to move cases around to keep pace with lateral partner hires). There will be synergies at work too.
The top 10 of the GAR 30 now looks similar to some peer-review-based rankings. For example there's a weak correlation between the firms in the top 10 and The International Who's Who of Commercial Arbitration (see table). For firms that don't do especially well in those peer review surveys, the GAR 30 can be a particularly worthwhile exercise. Two firms that get to show the vigour of their practices more in a data-led survey are Hogan Lovells and Baker & McKenzie.
In the table of Who's Who cases they appear outside the top 15 firms and in some other organisations' rankings they don't figure prominently. But in the 30, Hogan Lovells finishes 6th, and Baker & McKenzie 9th. Indeed, Hogan Lovells (especially the Lovells part) has consistently performed well in the 30. A year ago, the newly created firm was 3rd.
Baker & McKenzie, a year ago, was 23rd. Its rise appears related more to improved global practice management and communication than anything else - which is a fancy way of saying it's gotten better at counting arbitrations. (One of the revelations of the GAR 30 has been how few law firms have a clear picture at any point in time of how many live arbitrations they have on.) A few years ago, Baker & McKenzie brought in an ex-ICC counsel as global practice coordinator, to knit its various elements together better. In the process, the central record keeping function has improved. The result is a position 14 places higher than a year ago.
Still, if the top of the table now resembles more closely what one might intuit, there's little sign it's getting more competitive. No real challenger has emerged for Freshfields' crown in four years. If anything, it was being pushed harder in other editions of the 30 than it is today. On the really, really serious cases, Shearman & Sterling is clearly every bit of a match for it (as no doubt are some others). But Shearman looks unlikely to be able carry a workload equivalent to 40 merits hearings any time soon (It has 85 pending cases to Freshfields' 220). The WilmerHale practice has also, it's understood, been going through a period of reflection about what type of practice it should evolve into and particularly how far it wants to let its portfolio become eked out with non-premium work (as in work that comes with a price-cap attached). It's always a judgement call for any practice how far it can go towards taking work that will have to be done on a budget, with occasional corners cut, when so much about being good in this field depends on a reputation for flawless work product. You can't let the most senior arbitrators associate you with less than immaculate work.
Of the immediate chasing pack, only White & Case appears to have the partner infrastructure - a big enough team, divided across many different locations - to carry that caseload (and nearly is already) while also having individuals with the stature to pitch for the biggest assignments. White & Case has also been demonstrating more ambition of late. It now sponsors the leading scientific survey on international arbitration - the Queen Mary Survey - and has brought in a global practice manager: a lawyer poached from Allen & Overy to operate from New York. In the past, the word on the street had it that White & Case wasn't as unified as it projected itself, and at times different cities would pitch for the same work. Those sorts of stories haven't been heard for a while, and new leadership and stronger central practice management are only going to make things tighter still. Perhaps future editions of the 30 will show that it was around 2010 that White & Case ignited its growth. Similarly, as noted above, King & Spalding and Allen & Overy have been heading steadily upwards in the ranking for a number of years. Both are perceived by others in the market as the practices on the move.
Meanwhile, a story started doing the rounds a while back suggested Freshfields' is ever so slightly vulnerable. About 18 months ago, the Paris team was replaced mid-case by a long-standing client and Shearman & Sterling took over the work. So far it's apparently an isolated incident, but nevertheless it shows that nothing should be taken for granted. This is a business built on reputations and those can change reasonably fast.
Change has also taken place in the 30's middle order, arguably more so than in the top 10. Dropping into the middle order from last year's top 10 are CMS and Clifford Chance (last year's 6 and 4, respectively), while springing up from below are DLA Piper and Covington & Burling. Meanwhile Salans, Schellenberg & Wittmer, Jones Day and Baker Botts that featured mid-table are found lower down.
In the context of specific developments, those shifts aren't so surprising.
Covington & Burling - which has long had a strong brand for arbitration with US clients (remember it was the firm selected by Exxon to take on Venezuela) - has in the past yo-yo'd a bit in the ranking. But in recent years it's been making high quality lateral hires in Europe - from White & Case and Debevoise & Plimpton. That European team is now contributing work that counts towards the ranking score.
Clifford Chance's slip, meanwhile, has been on the horizon for some time. A series of developments - some positive - have slimmed down its practice. Those include the poaching of two senior figures by the ICC's Court of International Arbitration, coupled with - less helpfully - departures in the wake of the financial crisis. As one Clifford Chance partner recently quipped to a reporter, "our US litigation team is doing great. It's just working for someone else." Clifford Chance's core arbitration team, led by Audley Sheppard, retains its reputation as a formidable force.
CMS's slip, on the other hand, has a less clear-cut explanation. This is that organisation's third year participating, and finds it in 16th slot (its position of two years ago). In between it travelled up to 6th. Torsten Lörcher, the head of its international arbitration group, says that 2009 saw a significant increase of case numbers, which was at least in part attributable to the financial crisis. "A number of those cases", however, "did not go very far," he notes. Since then, the pace of merits hearings has fallen slightly - "from an increasing tendency of the parties to settle before a hearing takes place" while at the same time the sums in dispute have gone up -"significantly". Meanwhile, the GAR 100 researchers noted that when assessing the CMS portfolio, a lot of its work takes place under the rules of arbitral bodies in the former Soviet states, which aren't yet included in the basket of "sexy-rules" (to give them their official office title). So a significant proportion of its work goes uncounted for ranking purposes. CMS also suffered in the ranking as a result of another accounting change introduced this edition, on the credit given in the formula to arbitral appointments (CMS lawyers increased the amount they're sitting as arbitrators on an edition ago). And that's where we go next.
In the last edition of the GAR 30, the number of times members of a firm have been appointed arbitrator could count towards the final place. That favoured some of the smaller European firms and disfavoured larger international practices. Many of those put a cap on how often their partners are allowed to sit. You can make a case for setting the formula up either way: work as an arbitrator is invaluable (everybody agrees) to an advocate. But it does mean the playing field isn't level, for the larger firms.
This year we've opted to minimise the contribution arbitrator appointments makes to the GAR 30. We'll be doing a separate ranking on that which will also assess firms according to how highly their people are revered and how many cases they've collectively experienced, later in the year.
As a consequence, some of the civil law practices have fallen away. Mannheimer Swartling, with a volume of hearings way ahead of some international rivals (25 hearings made the grade), is now the highest placed firm from the civil law world. It went from 15 merits hearings a year ago to an impressive 25 this time. That was one of the biggest year-on-year rises. All that separates it from the top 10, it seems, are the parts of the formula that look at the value at stake.
Lalive, meanwhile, becomes the highest-placed Swiss firm. It's one of the few Swiss firms found working on its own mega-cases. It's also a little ahead of some in the market when it comes to advising on treaty-work. More Swiss firms now have some ability to do that, but Lalive has already filed pending cases. So the changes of emphasis in the formula work in its favour. Other firms who benefited for the same reasons are Latham & Watkins and Eversheds.
An extra word on Lalive. As in previous editions, the researchers had to make a difficult decision with Lalive's ranking. It's known through various channels that the firm is working on at least two state-to-state disputes of some magnitude - as in, sums that dwarf even the Yukos matter. However, the firm declines to supply specifics. Until our researchers are able to verify the details and the dates of the hearings etc, they're unable to factor the work in to Lalive's score. It's clear though that if two cases of that size were included, Lalive would be pushed into the top 10.
Sometimes the changes have an even simpler explanation. DLA Piper re-entered the ranking because this year it responded comprehensively to our inquiries (after taking a year out in 2009-2010). Another well known US firm, on the other hand, decided to take one of those year's out - ie, didn't provide enough about casework to go into the mix, having provided full information a year ago. (It's not named to spare any blushes.) Whether that's deliberate, researchers don't yet know. It's certainly hard, if you're a practice manager, to guess exactly where you'll finish in the 30 year on year - especially if you have one of the practices prone to yo-yoing. So it's not the most comfortable exercise for firms who like to micromanage their image. Still, as the great Wayne Gretzky said: "You miss 100 per cent of the shots you don't take." So every firm that takes part deserves enormous credit for putting themselves on the line. In the process, they help to render the world of arbitration more transparent. The next GAR 30 will be out in early 2012. Before then, look out for our ranking of practices according to their human capital, which will be on the website later in the year.
Both tables use T-scores to compare a firm's performance against the average performance for a particular column.
• The research period for all columns was 1 August 2008 to 1 August 2010.
• Column one shows the number of merits hearings participated in under "major arbitral rules". For this edition, the basket of approved rules expanded to include most Western European institutional arbitration rules (ICC, LCIA, DIS, Madrid Court, NAI, SCC, Swiss Rules, VIAC rules, Milan Chamber) along with matters under ICSID (including its additional facility), Cairo Regional Chamber, DIAC, HKIAC, Korean Chamber, SIAC, UNCITRAL and PCA administered. The column excludes any work other than final determination of liability (eg, jurisdictional hearings, hearings on document production, on settlements, quantum). Each matter is deemed to have a single merits hearing. If the lawyer who acted as lead counsel for the hearing is now at another law firm, then the merits hearing in question was assigned to the current firm. The column makes no distinction between work as co-counsel and work as lead counsel.
- Column three shows the number of merits hearings in the same time frame where more than US$900 million was at issue under any arbitral rules (in contrast to column one). It includes work as co-counsel and excludes matters not given a value or about declarative relief. Again, if the lawyer chiefly responsible in the hearing has since moved, as with column one, the matter is usually assigned to his or her current firm. Both sides to the case had to agree - broadly - on the value of a claim for it to score in this column. If arbitrators have handed down an award that is less than the amount being claimed, we have taken the arbitrators' award as the value.
- Column four shows the same information as column one, but in monetary value. It is the sum total of claims in the firm's merits hearings.
- Column five shows the median value of the firm's merits hearings, in dollars.
- Column six shows the number of pending arbitrations on a firm's books.
- Column seven shows the number of cases from column five that are occurring under investment treaties.
- Column eight shows the dollar value of all pending cases in which a firm is counsel. It omits anything still at the pre-action stage.
- A score is calculated by compiling the sum of performances in all columns - using T-scores and a universal factor (worth three figures). The score is not published in this year's table.
Q&A on methodology
What's the source for this data?
All figures used in the GAR 30 are supplied by the firms who want to be in the survey (96 sent in data). To count, each matter must be accompanied by sufficient information that can be verified (for example, date, duration, value, name of opponent, industry affected, names of arbitrators or chair, etc). If in doubt, we leave it out.
How did US$900 million come to be the cut-off for column three?
We've been asked to ensure greater credit is given to firms that are sought out on high-value claims. Hence there are now two columns that filter the amount of value at stake in a firm's completed hearings. One of them uses the figure $900 million as the threshold. It is an arbitrary cut-off point, simply designed to find "very high value"work.
Do you accept the values asserted in ICSID claims at face value?
Yes. But, with claims of all stripe, once there is a known award, that award becomes the value of the case for our purposes.
Why don't you include ICSID jurisdictional hearings?
The same question is now appearing for annulment work. The position on both is under review. For now, ICSID work remains the junior partner in the arbitration market. Therefore we are not emphasising it.
We had more matters in the $900 million range than you have counted - why?
We were particularly cautious in accepting matters for column two. A number of issues arose:
(i) the law firms on either side disagreed about the value at stake;
(ii) no value was given other than a statement that it was "worth billions";
(iii) requests for declarative relief;
(iv) baskets of Public International Law matters before unique claims tribunals;
(v) test cases;
(vi) hearings "happening soon"; and
(vii) acting as co-counsel.
With issues (i) and (ii), we made further inquiries, or went with the lower value. With requests for declarative relief (iii), we assessed items individually. It is fair to say that, in general, a basket of public international law matters to be materially different from a "bet the company" case, partly because of the volume aspect to the work and partly because the tolerances for a state and a company are different. Nevertheless, we consider each proposal individually and a few have in the past been accomodated. With (v) test cases, we took the same approach as with declarative relief. As for (vi) hearings due soon, the item fell outside our research window and cannot be counted until next year's survey. We did count matters where a firm was acting as local co-counsel (vii). Throughout, we have tried to use common sense and caution. A general statement of our approach would be, as before: if in doubt, leave it out.
These figures seem entirely self-reported. How do I know that they're accurate?
We require firms to supply sufficient information about each hearing that can be verified.
Did every firm manage to supply every figure you asked?
Pretty much, yes. The only column where some are reluctant is the value of all claims in the portfolio, for various reasons. However, some firms are poor at specifiying the date and duration of a particular hearing. This counts against them, as if there are too many blanks in the table alongside a particular matter then it doesn't get counted. If in doubt, leave it out.
Why do you weight some columns in the score?
Left to our own devices we would print the table strictly according to the two columns on merits hearings. That has proved too rich for some firms' taste, so we have created the formula described in the analysis accompanying the GAR 30 table, since the raw scores for each column are also given and readers are free to sort the firms using their own recipe.
Lots of my cases settle - do you take the rate of settlement into account?
At present - no. The GAR 30 regards settlement as affecting all practices equally. International arbitrations do often settle, but it is less common than in litigation. In addition, it would be curious for all the partners in supposedly big and active arbitration practice to experience a slump in merits hearings at the same time.
Z-scores and T-scores
A Z-score, also known as a standard score, measures a performance in a test against the average score achieved, expressing the gap between the two in terms of the standard deviation. In English: it means you can focus on how well a person or organisation did relative to others and also give credit for doing well in harder tests. For example if a student scores 63 per cent in a test where others score no higher than 40 per cent and then 80 per cent in a test where most score the same, it's important that the more impressive of the two performances (in relative terms) get credit. But how? Merely adding percentages would drown it out. T-scores avoid that problem by putting everything into the same scale, so it carries equal weight.