Marc J. Goldstein Arbitrator & Mediator NYC
August 07, 2017

Salvation of the Baby-Splitters

Perhaps one should have a valid excuse, in the company of intended readers of Arbitration Commentaries, to re-open the well-worn subject of compromise outcomes on damages in international arbitration – “triangulation,” in the parlance of some eloquent and perceptive commentators (e.g., M. Kantor, Avoiding Triangulations and Chimeras Alike, Global Arbitration Review (July 31, 2012))*; but here identified by its colloquial name: “baby-splitting.”  I claim to have such an excuse, having recently been permitted to participate as a speaker in a program about damages before a knowledgeable audience, and for that purpose having tried to compose some thoughts on the matter. A few of those thoughts, falling short (as these essays usually do) of being a definitive treatment, are set forth here:

  1. “Institutional/environmental” factors often weigh against baby-splitting (at least overt forms of it). Let’s first consider investor-State arbitration. Investment arbitration awards, unless made confidential by agreement or order, are published in accessible online sources. They are studied and discussed by practitioners in the field and scholars in academia. The transparency factor is one element that should encourage investment arbitrators to be thorough and objective in their approach to damages. The relatively concentrated nature of the investment arbitration bar, and its well-developed communication networks, also serve to place investment arbitrators under a microscope. The prospect of annulment proceedings in ICSID cases is perhaps a factor – but high professionalism may be the best explanation – that motivates investment law arbitrators to display in their awards painstakingly detailed and thorough exposition of the parties’ contentions and equally painstaking Tribunal analysis on both liability and quantum. This would seem to be not particularly out of fear of an actual annulment, but for two reasons: first, to inhibit the prospects of and the costs associated with annulment proceedings, and second, to avoid the negative peer review (with full transparency to the investment arbitration community) that could appear in an annulment committee’s decision even if the award is sustained. Let’s now consider commercial arbitrations, with lesser degree of institutionalized transparency. In ICC arbitrations and others where there is similar “peer review” of draft awards (members of the ICC Court who scrutinize awards being viewed as peers of the arbitrators for this purpose), the incentives to resolve quantum issues analytically are also present. Many arbitrators approach damages analytically mainly because this is what arbitrators consider that they are duty-bound to do. Whether subjective choices of damages amounts are more prevalent where the awards are less scrutinized and less transparent is difficult to say. I know many arbitrators who consider themselves duty-bound to tackle difficult quantum issues as objectively and thoroughly as the evidence and circumstances permit, without regard to transparency factors. At the same time, we know that wrestling with complex quantum issues demands great discipline, stamina, and time, and presiding arbitrators are sometimes selected for reasons having little to do with their willingness or ability or availability to make the required effort. Some of them will compensate by delegation, whether to another member of the Tribunal, or to a Tribunal assistant, who claims to have quantitative skills and a consequent ability to decipher and analyze the evidence about damages.
  2. Are Disappointed Parties Propagators of the Myth of Baby-Splitting?: I have no survey data to support this hypothesis. But is seems entirely plausible to suppose that some of what is perceived as “baby-splitting” by arbitration participants is a speculative attempt to ascribe a subjective motive to the arbitrators’ rejection or partial rejection of the party’s position on damages. When arbitrators have latitude to omit detailed statements of their reasons – as they do in many settings other than investment law cases, the arbitrators may have specific reasons for omitting their reasons. For example, an award that is highly critical of the analysis or the credibility of an expert to whom a party has presumably paid considerable sums in professional fees may strain relationships. Arbitrators may prefer to leave a gap in the reasoning, seeing reticence as the lesser evil. In such a case, the perception of baby-splitting may be justified, but the reality may be otherwise. Conversely, arbitrators may state detailed reasons for having adopted particular assumptions or discredited particular elements of the Claimant’s position on damages. And yet the Claimant and its counsel may remain convinced that this rejection was subjectively not analytically motivated, the award’s stated reasons being seen to operate as an arbitrary tool to make a subjective outcome appear rational. These kinds of reactions are understandable byproducts of an adversary process that entails substantial financial and emotional investments by the parties and their professionals. Consider this example. In a recent investment arbitration case, Claimant claimed that the expropriated assets had a 40-year expected life, while the Respondent State contended for 25 years. Each party submitted extensive expert evidence on this specific question. Opting for the 25-year term, the Tribunal provided detailed assessments of the reliability and credibility of each expert’s submissions. This choice had a 35 percent reduction impact on valuation under the discounted cash flow (DCF) methodology. Will the Claimant or its professionals admit, in any public forum or private survey, that Respondent had the better of the argument? Perhaps not. But if the award is the objective evidence, and the arbitrators’ motives and stated conclusions are presumed to be sincere, the case for baby-splitting is difficult to make.
  3. Survey Evidence Disputes Baby-Splitting as Mostly Myth: Evidence that arbitrators do not systematically gravitate toward moderation in damages outcomes comes from some sources even more authoritative than your Commentator. A 2015 study by the American Arbitration Association covering 2,384 AAA domestic and international commercial arbitration awards in cases where monetary relief was claimed found that 41 percent of the cases resulted in awards of 80 percent or more of the amount claimed, 31 percent resulted in claims denied, and only 28 percent resulted in awards of less than 80 percent of the amount claimed. (“AAA/ICDR Awards Do Not Split the Baby: Countering Counsel Perception in Commercial B2B Arbitration Cases,” a 2016 AAA report that may be found at; for apparently technical reasons owing to changes in the AAA’s website, this AAA whitepaper is not currently posted there). A broader study in 2014 that canvassed multiple empirical analyses concluded: “Practically all the research under consideration came to the conclusion that arbitral tribunals very often grant claimant or respondent the full or almost the full amount claimed; awards in the mid-range are the exception and not the rule.” (A.C. Weber, C.A. Pascuzzo & G. De Siqueria Pastore, Challenging the “Splitting the Baby” Myth in International Arbitration, 31(6) Journal of International Arbitration 719-734 at 733 (2014)). Such surveys have their own limitations of course. A survey of all AAA/ICDR awards in a given year may include a large number of cases where the amount of the award, if there is liability, is relatively well-defined by contract. A more selective sampling of business valuation and lost profits cases would probably yield different results.
  4. Salvation of the Baby-Splitters?: Even if there is no consensus about how prevalent baby-splitting may be, we may all agree that it would be desirable to take concrete steps to reduce the arbitrator’s temptation. Here there is shared responsibility of advocates and arbitrators. Suppose, as is wont to happen in some cases, the Claimant submits a thorough DCF analysis with a specific outcome, and Respondent’s expert submits an “attack” report that questions Claimant’s experts’ assumptions, methodology, data accuracy, etc., but does not make a counter valuation or even a sensitivity analysis on key variables (or, mainly for rhetorical purposes, opines that there are no damages, that the value of the property taken is zero or negative, etc). Responses of arbitrators in such situations may in part be cultural. For American arbitrators, especially in a case with American counsel, the norm of party autonomy translates into a strong momentum toward allowing the parties to present their cases as they wish. If this leaves the Tribunal at the deliberation stage with less guidance on quantum than might be desirable, at least the dysfunction is a byproduct of the parties’ choices. But even American arbitrators do find ways to manage the problem. One approach, suitable for a case where quantum stands to be a complex valuation or lost profits endeavor with foreseeably conflicting expert testimony and substantial potential variations based on data and assumption choices, is to encourage the Respondent, even at the preliminary conference stage, to respond to Claimant’s quantum evidence with a quantified counter-position. Another possibility is to await receipt of both sides’ expert submissions, and if Respondent has made an attack report only, to invite this to be supplemented with a counter-calculation. (More commonly, but perhaps less effectively, the Claimant submits a reply expert report purporting to show that Respondent’s criticisms have been taken into account, resulting in modest adjustments).  Another strategy is for the parties’ experts to share their damages models with the Tribunal, or to develop jointly a serviceable model for the Tribunal’s own use. (This phenomenon can be observed in the reading of a recent investment arbitration award, see Burlington Resources v. Republic of Ecuador, ICSID Case No. ARB/08/5, Decision on Reconsideration and Award dated Feb. 7, 2017, published on This solution raises further issues about the Tribunal’s ability to manipulate the model, the possible need for a Tribunal-appointed expert, and the possibility of ongoing dialogue with the party experts beyond the confines of a merits hearing.


*“Triangulation” in social science terminology refers to the use of multiple methods in a study in order the cross-check the validity of the results. In the arbitration context the term initially had reference to an arbitral tribunal’s examination of possibly three different quantum outcomes based on possibly three different valuation methods  — income-based (DCF), market-based (comparables) and asset-based (net asset value).  See C. Amirfar, The Role of Damages Calculations in the Legitimacy of International Investment Arbitration, in A. Rovine, ed., CONTEMPORARY ISSUES IN INTERNATIONAL ARBITRATION AND MEDIATION: THE FORDHAM PAPERS 2015, Chapter 7 at 108, 110 & nn. 10, 11, citing to M. Kantor, supra.)

Leave a Reply