Mr Daniel Kahneman is a Nobel Prize winner in Economic Sciences, and the author of the bestselling book “Thinking, Fast and Slow”. His book focuses on behavioural science, and explains how cognitive biases fool us into making suboptimal decisions. In December 2017, PwC updated its International Arbitration damages research (“PwC Research”). It reviewed multiple international arbitration proceedings with the view of establishing a correlation between the value of damages calculated by parties’ experts and the value of damages actually awarded by tribunals.
Critical questions arise when we compare Mr Kahneman’s work with the PwC Research’s findings. In particular, questions relating to whether parties in international arbitration are able to influence the value of damages in awards, by (mis)leading arbitrators into cognitive biases.
Mr Kahneman describes “anchoring effect” as a cognitive bias phenomenon. It occurs when a person is asked to consider a particular initial value, relating to an unknown quantity, before estimating that quantity. What follows is that the person’s estimate tends to remain close to that value initially considered; even in situations where the latter bears no correlation with the former.
Any number we are asked to consider as a possible solution to an estimation problem will induce an anchoring effect. Mr Kahneman illustrates the power of anchors through an experiment conducted with German judges with an average of 15 years of experience. The judges reviewed a case relating to a woman who had been arrested for shoplifting. The judges were then asked to roll a rigged dice, which would stop at either three or nine. Right after, the judges were asked whether they would sentence the woman for a term in jail longer or shorter, in months, than the number showing on the dice. The judges were then instructed to hand down the sentence to the shoplifter. On average, those who rolled a nine would sentence her to eight months; those who rolled three would sentence her to five months. The anchoring effect in this case translated into a difference in jail time of more than 50% – even though the dices were obvious random anchors.
Compatible results have been found in experiments with other experienced professionals asked to consider an initial value; such as real estate agents asked to estimate the price of houses, and investment fund managers asked to estimate the price of companies’ shares.
It turns out the anchoring effect is one of the most reliable phenomena in experimental psychology. And anchors that are obviously random can be just as effective as potentially informative anchors.
On the basis the anchoring effect has been proved by scientific research, the questions that follow in international arbitration are:
(i) Are claimants incentivised to set out an inflated amount for damages at the outset of the arbitration, in order to anchor tribunals? and;
(ii) Should respondents assume that claimants usually attempt to anchor tribunals and, for that reason, equally attempt to anchor tribunals on the lower side of the spectrum?
PwC Research reviewed 116 publicly available awards. It draws a correlation between the value of the damages claims calculated by parties’ appointed experts, and the value of damages actually awarded by tribunals (in the cases where claimants obtained a favourable award). The PwC Research is interesting in a number of fronts. For the purpose of this post, the relevant findings are:
(iii) Tribunals awarded on average 36% of the value of damages calculated by claimants’ experts.
(iv) Respondents’ experts on average assess a claim at 12% of the value calculated by claimants’ experts.
(v) In situations where respondents’ positions move closer to the claim value calculated by claimants’ experts, the tribunal’s award does the same.
How is it possible that skilled and impartial experts calculate values so far apart? The PwC research rightly explains that often (a) experts answer to different questions; (b) experts are instructed to treat facts differently; and (c) experts genuinely have different opinions.
But even where tribunals instruct experts to consider the same legal and factual assumptions in their calculations; in many cases, experts find significantly distant values. This is caused by, amongst other things, the application of different (seemingly valid) complex mathematical models, which tribunals may or may not understand.
In any event, should it be acceptable to leave tribunals to their own devices? As the PwC Research shows that tribunals have been asked to consider values with a delta of 88% between claimants and respondents’ experts’ damages calculations?
Moreover, does point (iii) above suggest that claimants have inflated their claims? Does point (v) above suggest that respondents have an incentive to depart wildly from the value of damages put forward by claimants?
Comparing Mr Kahneman’s work with the PwC Research’s findings raises difficult questions. In international arbitration, could parties influence the conclusions of tribunals by anchoring strategies?
If the answer is yes, there is (arguably) a risk of international arbitration becoming a dysfunctional process. Should tribunals, therefore, seek to deter anchoring strategies? If yes, should tribunals consider such strategies as parties’ misconduct?
One could argue that tribunal’s power to allocate the costs of the arbitration should per se discourage parties from deploying anchoring strategies. For example, if claimant does exaggerate the value of its damages claim, with the view of anchoring the tribunal; even if claimant succeeds in the arbitration, tribunals are (usually) empowered to allocate costs according to the proportion of claimant’s success.
That argument is potentially flawed, because tribunals generally pay particular attention to the outcome of the arbitration when allocating costs. If the outcome of the arbitration was tainted by anchoring strategies (i.e., the damages awarded were based on a distorted valuation); then the basis for allocating costs would be artificial.
Alternatively, should tribunals focus exclusively on parties’ conduct to allocate the costs of the arbitration? Besides denying the recovery of the arbitration costs, tribunals may order successful claimants – that have deployed anchoring strategies – to pay the costs of the defeated party on an indemnity basis (e.g., ICC Rules and LCIA Rules empower tribunals to do so, if the arbitration agreement does not provide otherwise).
That may also be insufficient to deter anchoring strategies. Parties willing to adopt such strategies could calculate the cost of the potential penalty (i.e., not recovering its costs and even paying the costs of the other side); and compare it with the potential reward (i.e., anchoring the tribunal on value substantially higher or lower than the actual value of damages).
If the risk does not outweigh the reward, the party deploying the strategy could take a commercial decision on the basis of a cost benefit analysis.
Lastly, the most difficult question: how can tribunals detect anchoring strategies deployed by parties in international arbitration?
While scientific research has proven the anchoring effect, many will be confident that they understand such phenomenon, and can temper their decisions accordingly. This is caused by another cognitive bias namely “an illusion of explanatory depth”; which deserves a separate post.
This post advances more questions than answers, and hopes to raise the debate as to whether cognitive biases have critical implications for international arbitration.
Compelling indications, especially after the PwC Research, suggest that anchoring strategies could influence the value of an award on damages (at least for the time being, before human arbitrators are substituted by robots).
The invitation is out for the international arbitration community to discuss whether anchoring effect could (unduly) impact the conclusions of a tribunal. If we are prepared to accept that premise, we should then proceed to the second step and propose measures to counteract cognitive biases – with the overriding aim of increasing consistency in international arbitration.