Surveys conducted by Queen Mary University of London have confirmed that international arbitration remains the preferred method for resolving cross border disputes. One of the key attractions of arbitration is the ability of the parties to participate in the selection of the arbitrators who will determine the issues in dispute and the quality of arbitration proceedings is largely dependent on the quality and skill of the arbitrators appointed.
Construction and energy are two sectors where the complexity and technical nature of disputes favours a process where the parties can select the person who will decide the claims. When disputes arise in the construction and energy industries, the amounts at stake can be substantial and the arbitrators are frequently required to make complex decisions, particularly on the question of damages. The confidence of parties in the arbitration process very much depends on arbitrators being able to make well-explained and financially sound determinations of damages.
Recent cases, notably the Yukos arbitrations, have highlighted how unpredictable the assessment of damages can be and the impact that this can have on the perceived legitimacy of the arbitral process, particularly in the context of investor-state arbitration.
In 2015, PWC published the results of a research project into the assessment of damages in international arbitration cases, analysing 95 publically available awards. The report examined a number of issues, including the role of experts and differing valuation approaches applied by tribunals. One of the other points considered in the report was the extent to which cognitive biases, in particular “anchoring” and “extremeness aversion” may influence the tribunal’s thinking on numbers.
What is cognitive bias?
The concept of cognitive bias is probably best explained by reference to Daniel Kahneman’s bestseller “Thinking, Fast and Slow”, which summarises decades of research into how we make decisions. The book’s central thesis is the dichotomy between two systems of thought:
- System 1: fast, automatic, based on intuition.
- System 2: slow, effortful, based on deliberation.
The human brain, Kahneman explains, automatically and subconsciously uses mental short cuts (heuristics) to make complex decisions. These mental short cuts allow us to operate in a world that increasingly bombards us with information but they can lead to errors (cognitive biases) which can, in turn, result in mistakes. Examples of cognitive bias that may affect decision-making are “anchoring” and “extremeness aversion” (also known as “splitting the baby”).
When estimating a numerical amount, people usually start with an initial value, an anchor, and then come up with a final estimate by making adjustments to the anchor.
For example, in the Yukos arbitrations, the tribunal started on the basis of the numbers submitted by the claimants and adjusted them downwards, taking into account some of the respondent’s criticisms and its own considerations with regard to some of the assumptions.
If the anchor provides useful information about the underlying value and if the adjustments are reasonable, anchoring can be a useful way of reaching a decision. The problems arise where the anchor is irrelevant and the adjustments are unreasonable.
The PWC research showed a degree of correlation between a tribunal’s award on damages and where the parties position themselves on damages. As respondents’ positions move closer to the claim value, the tribunal’s award does the same. The findings suggest that arbitrators may be susceptible to anchoring and, for parties, it highlights the importance of thinking carefully about the numbers used to estimate damages and any adjustments that need to be made.
Individuals are inherently averse to extremes and tend to compromise between extremes, even when the extremes are irrelevant. In arbitration, this gives rise to the perception that arbitrators are unwilling to rule strongly in favour of one party and adopt a form of extremeness aversion by “splitting the baby”.
In fact, the empirical evidence suggests that “splitting the baby” is a perception as opposed to a reality in international arbitration. The results of the PWC research showed no clear trend of tribunals going for the middle ground when awarding damages. This is very much in line with the results of a US study conducted by Richard Naimark and Stephanie Keer based on awards made through the American Arbitration Association during the years 1995-2000 (Stephanie E. Keer, Richard W. Naimark, ‘Arbitrators Do Not “Split the Baby” Empirical Evidence from International Business Arbitrations’ (2001) 18 Journal of International Arbitration, Issue 5, pages 573–578.).
The PWC research also showed that in cases where the amount awarded appeared to be in the middle, it was often the result of the tribunal deciding on a number of issues (some in favour of the claimant and others in favour of the respondent) with the resulting award reflecting those decisions.
In spite of the empirical evidence to the contrary, concerns about tribunals “splitting the baby” persist. Respondents to the 2012 QMUL Arbitration survey considered that that tribunal unnecessarily “split the baby” in 17% of their cases. Such concerns tend to be greater in cases where tribunals do not provide clear or sufficient reasoning for the award, particularly with regard to the assessment of damages. Greater transparency in how tribunals determine the amount they award, including their determination on each of the key areas of difference can help dispel some of those concerns.
Most national laws and arbitration rules require arbitrators to state the reasons on which awards are based, so that the parties can understand the rationale behind the arbitrator’s decision. In complex disputes, the determination of damages is often a highly sophisticated exercise, involving financial analysis and economic modelling, which can make it difficult for a tribunal to articulate the criteria applied in order to arrive at a particular figure. In spite of this, PWC’s research findings indicate that arbitrators are becoming increasingly sophisticated in their assessment of damages; explaining their approach to damages in more depth and addressing more complex valuation issues, than was historically the case. This tends to suggest that arbitrators are increasingly resistant to the types of cognitive bias that can cause us to think fast instead of slow.