ICSID Arbitrators: The Ultimate Social Network?

Michael Waibel is a University Lecturer at the University of Cambridge.

Sergio Puig’s article offers a refreshingly new, thought-provoking analysis of the links between investment arbitrators. Who the parties appoint as arbitrators matters for how the investment treaty regime operates. Criticism of the appointment process in investment arbitration is widespread, yet rigorous empirical work on this important aspect of the investment treaty regime remains rare. Relying on the theoretical frame of social network analysis, Sergio sheds light on the interactions among ICSID arbitrators. His analysis suggests that a core group of 25 arbitrators enjoys disproportionate influence on the development of the investment treaty regime. To understand what animates the regime, it may suffice to analyse the preferences and political philosophy of these 25 ‘power-brokers’.

The Core-Periphery Divide and Underrepresentation

The core of the arbitration network is composed of 25 arbitrators. Members of this elite group are on average connected to 11.75 other arbitrators. The core’s cultural and legal homogeneity is striking: 14 Europeans, 4 Latin Americans, 3 Canadians, 1 New Zealander and only 1 US arbitrator. Similarly, among the top ten countries of origin of ICSID arbitrators overall, five are Euro-Atlantic states (US, France, UK, Canada and Switzerland) and five Latin American. By contrast, African and Asian arbitrators have rarely been appointed to ICSID tribunals, despite significant inward- and outward flows of foreign investment to and from Asia in particular. In such data-intensive work, minor errors are bound to creep in. For instance, in Figure 4, Vaughan Lowe appears as a US, rather than a UK national, and the 26^th arbitrator in the core between Jonny Veeder and Jan Paulsson remains nameless.

The formal bond of nationality may be only a crude measure of arbitrator behaviour. As Sergio rightly highlights (p. 405), many arbitrators with nationalities of developing countries have received at least part of their legal education in developed countries (chiefly the UK, the US and France). Indeed, if one focuses on arbitrators who have spent most or all of their adult lives in a developing country and have not worked or received part of their educated in developed countries, the voice of developing countries in even less represented among ICSID arbitrators than the formal link of nationality would suggest.

Other demographics are also underrepresented among ICSID arbitrators, most notably women. 93 percent of all ICSID appointments were male arbitrators (p. 404). Given that Brigitte Stern has accumulated the highest number of appointments of any ICSID arbitrator, the share of female arbitrators is even lower at 5 percent (p. 405). Two female super arbitrators apart, women are at the periphery of the arbitration network. This gender imbalance mirrors the general characteristics of the entire arbitration network (p. 411).

Yanhui Wu and I have recently assembled data on a control group composed of more than 700 potential ICSID arbitrators, i.e. individuals with similar characteristics and qualifications to those who have already been appointed to at least one ICSID tribunal. Our control group includes current and former ICJ and WTO Appellate Body (AB) Members who, unlike some of their judicial colleagues, have yet to be appointed to an ICSID tribunal, and partners at leading arbitration practices in the same position. Ten former AB members and 18 current and former ICJ judges since 1990 have never been appointed. The following table compares some characteristics of the treatment and control groups.

The comparison of the treatment and control groups shows that arbitrators in the control group are on average 6 years younger than their appointed counterparts, and less likely to be in full-time private practice. They are also more likely to be female, nationals of developing countries and specialists in public law. Graduates of elite universities are overrepresented among appointed arbitrators. As Table 1 in Sergio’s paper also shows, French arbitrators feature very prominently, surpassed only by the United States. Moreover, Sergio notes that the high absolute number of US arbitrators masks that ‘most Americans … remain at the periphery of the network’ (p. 410) because the majority are only ever appointed to a single tribunal. Overall, those with a common law legal origin are underrepresented relative to those with French legal origin.

Barriers to Entry

The barriers to entry for those aspiring to sit as arbitrators in investment treaty arbitration have traditionally been high. These small world properties could have important implications for how tribunals decide cases. Not only is the world of investment treaty arbitrators technocratic, it is also a highly unequal one within. Investment arbitrators are not created equal. The rise of superstar arbitrators (which Daphna Kapeliuk documented in her earlier work) has created a divide between one-time and repeat players. Most of the spoils go to what Sergio terms the ‘power brokers’ and the ‘elite arbitrators’, similar to the highly unequal distribution of economic payoffs we observe in elite sport, music or even among cult academics such as Steven Pinker or Michael Sandel. Sergio finds that 10 percent of repeat arbitrators account for 50 percent of all appointments, whereas more than 50 percent of all ICSID arbitrators are ‘single shooters.’

This skewed distribution reflects that both institutional and especially party appointers favour arbitrators with a track record of appointments. Despite investing considerable time and money in selecting the ‘right’ arbitrator, party appointers operate under incomplete information and may not be able to forecast how a newcomer would behave once appointed. Risk aversion and path dependency may play an important role in skewing the appointment process towards ‘insiders’. Through preferential attachment, arbitrators present at the rise of the ICSID regime – in the late 1980s and early 1990s – are much more likely to find themselves at the core of the network. They have a competitive advantage over newcomers. Experienced arbitrators are selected for a new dispute because other parties chose them previously.

Fellow arbitrators may have better information and existing ties of influence over insiders, and may prefer to appoint other experienced arbitrators rather than newcomers as president of the tribunal. Through this self-reinforcing dynamic, the influence of a small group of superstar arbitrators grew as the regime expanded and gained in salience. Is this a largely random process, or would the same individuals be at the core of the arbitration network if we went back in time and started from scratch? How important are the initial conditions and the timing for the emergence of particular individuals are super arbitrators? It is likely that there is a considerable unpredictability in these appointments.

Extensions

By introducing network analysis as a new methodological tool to international law, this extremely stimulating paper opens up many new avenues for enquiry, in investment arbitration and more generally. I focus on four.

First, I would also like to see further analysis of the distinction between institutional appointments and party-appointments (for the most part made by non-arbitrators), and the 73 appointments of presidents by party-appointed arbitrators by common agreement. These are the only appointments over which insiders have control, and they are thus likely to be particularly illuminating on what matters for such appointments.

What really explains these peer appointments? Sergio tells us about the role of ‘authorities’ and ‘hubs’ in Figures 6 and hints at a potential role for reciprocity, but stops short of analysing whether appointments are driven by (i) the appointee’s standing  among fellow arbitrators, (ii) reciprocity among arbitrators, or (iii) strategic choices by party-appointed arbitrators. As to (ii), party-appointed arbitrators might seek appoint as Presidents individuals whom they might be able to convince come decision times and form a majority with. With all three mechanisms, the exclusion of outsiders is a by-product, rather than an objective of insiders.

Second, Figure 6 and Table 3 in the paper do not appear to be entirely consistent with Figure 4. For example, the Philippine jurist Florentino Feliciano comes highest in a combined hub-and authority score, yet does not appear in the core 25. Sergio might want to comment further on how figures 4 and 6 compare as measures of the centrality and importance of arbitrators to the regime, and on which measure of centrality is the most meaningful in this particular context.

Third, how the financial incentives for arbitrators affect the operation of the regime deserves further study. Sergio states that ‘arbitrators make on average $200,000 per case’ (p. 398), but does not explain how this average pay-off is computed. An extreme example is an unusually complex and lengthy (PCA) arbitration, Yukos v. Russian Federation. The average compensation for each of the three arbitrators was EUR 1.75 million (US$ 2.25 million). Our understanding of how different types of potential and appointed ICSID arbitrators regard these levels of compensation is limited. For some lawyers with high incomes from their main activity, the opportunity cost of an ICSID arbitral appointment may be high. One of  Sergio’s interviewees went so far as to suggest that to some ICSID arbitration was a pro bono activity (fn. 61).

Fourth, while Sergio notes ‘intense clustering’ among ICSID arbitrators, he tells us little about how nodes may form particular communities within the investment regime. Drawing on Anthea Robert’s work on clashing paradigms, it appears particularly worthwhile to analyse the role of different epistemic communities within the investment regime, based on expertise (public international law specialists vs. commercial arbitrators vs. public lawyers) and main occupation (academics vs. full-time private practitioners vs. government lawyers). Social network analysis promises new insights on the formation of distinct communities, and how they overlap.