Behavioral International Law: What Is It Good For?
Having set out general considerations and a research methodology framework for “Behavioral International Law” in previous posts, some readers might be wondering how this all cashes out for international law as a discipline?
In their path-breaking 1999 YJIL article on economic analysis of international law, Jeffrey Dunoff and Joel Trachtman noted that “almost every international law research subject could be illuminated, to some degree, by these research methods” [referring to economic analysis]. With similar caveats, it is tempting to say something similar and related about behavioral analysis and international law. Behavioral international law is not a ‘theory of everything’. Neither is it a normative framework of analysis, as such. But properly constructed behavioral research selectively employing the methodologies I describe here can significantly increase our knowledge in all areas of international law, with respect to many problems and puzzles.
In my article, I developed three examples that cover the entire spectrum of levels of analysis as well as research methodologies. In all of them, a mere theoretical application is sufficient to stimulate discussion by posing alternative hypotheses and explanations, but if one is concerned with empirical accuracy, field studies and experimental work is necessary. Moreover, the examples – essentially three mini-articles – cover diverse areas of international law (treaty law, WTO dispute settlement and international humanitarian law). I will briefly summarize two examples.
In the first example I engage with recent literature addressing the malaise in effective treaty-making in some important areas of international law. Specifically, in this VJIL article Andrew Guzman applies economic analysis to treaty law, ultimately calling for a dilution of the consent requirement in international treaty-making. As Guzman explains, a Pareto-efficient improvement in treaty law (whether reforms to an existing treaty regime, or the launch of a new one) would entail that at least some parties gain from it, but none are made worse off by it (in comparison to the prevailing legal situation prior to the change). In contrast, a Kaldor-Hicks-efficient treaty improvement would be such that the benefit granted to some states from the change would be greater than the disadvantage caused to other states; in other words, aggregate utility is increased, although there are winners and loser in a distributive sense. To Guzman, the strict consent requirement in international law-making epitomizes a Pareto concept of efficiency; aggregate welfare is sacrificed to the maintenance of individual state welfare. Progress in international treaty-making in areas of pressing concern to international community as a whole, such as climate change, are blocked because the consent requirement empowers states to block it – which they do, because of their specific utility.
A behavioral approach could offer an alternative analysis that simultaneously strengthens Guzman’s critique and yet produces different policy prescriptions: the cognitive status quo bias. In the face of Coasean economics, cognitive psychology provides a startling insight: all things considered, people actually have a tendency to prefer an existing state of affairs over alternatives that might leave them better off. Experiments with individuals show that they disproportionately prefer the status quo over alternatives when making decisions. Do states (and/or elite decision-makers who determine state preferences) suffer from a similar bias?
Theoretically applied, the immediate implication would be that states might regularly be withholding consent from new treaty regimes and treaty reforms, not (or not only) because they believe these regimes and reforms run counter to their own best interests, but because of an embedded behavioral aversion to change. Thus, where Guzman sees the consent requirement as the cause for an empirically observed ‘status quo bias’ in international treaty law, it might actually be the case that the excessive commitment to consent that he decries is in fact the formal result of a behavioral status quo bias. In other words, states preserve the constraints of the consent requirement, because it enables them to pursue their (imperfectly rational) preference for status quo.
On the one hand, this makes Guzman’s descriptive analysis even stronger because it indicates that the consent requirement may even facilitate the scuttling of treaty regimes and reforms that would be Pareto efficient (i.e., when no state has a ‘rational’ reason to object), not just those that are Kaldor-Hicks efficient. Treaty reform can therefore be exceedingly difficult even when everybody stands to gain and no one stands to lose. On the other hand, the policy implications of this theoretical application (which in fact has some empirical grounding) are that relaxing the consent requirement would not be sufficient to solve the problem. In many cases, states will continue to eschew formal legal change – consider the WTO, where majority voting on treaty amendments, waivers and other decisions is formally permitted, but hardly ever used.
Skipping the extended analysis of conformity biases in judicial decision-making that bears on the debate over the role of dissent in international tribunals, at the other end of the methodological spectrum, the third example I develop relates to the rationality of military commanders or JAGs, as individuals or as part of small decision-making cells, in the assessment of proportionality in targeting decisions. Here I suggest that such decisions might be subject to cognitive ‘framing effects’. Experimentally proven framing effects arise when alternative descriptions of the same decision problem give rise to different preferences, often as expressions of the endowment effect. Thus, in one classic example, people who have lost 10$ on the way to the theater will tend to nonetheless purchase a ticket; but people who have lost a pre-purchased ticket to the theater worth 10$ will not, even though in expected utility terms, the rational decision should be the same.
Closely following the classical ‘Asian disease’ experiment conducted by Amos Tversky and Daniel Kahneman (The Framing of Decisions and the Psychology of Choice, 211(4481) Science 453 (1981)), I suggest the following general design of an experiment, which could be conducted of a variety of samples. Two groups of subjects would be presented with the following hypothetical:
Imagine that the enemy has stockpiled a significant quantity of strategic munitions in the basement of an apartment building that normally houses 200 civilians. As part of a military operation your unit has been assigned the mission of eliminating this stockpile. You have two alternative plans of action to choose from, both of which will eliminate the munitions.
A first group of subjects would be posed with this question:
If plan A – a manned aerial attack – is adopted, 80 civilians will survive. If plan B – an unmanned drone attack – is adopted, there is 1/3 probability that all civilians will survive, and 2/3 probability that no civilians will survive.
A second group of subjects would be posed with what is in essence the same scenario, but framed differently:
If plan A – a manned aerial attack – is adopted, 120 civilians will die. If plan B – an unmanned drone attack – is adopted, there is 2/3 probability that all civilians will die, and 1/3 probability that no civilians will die.
These decisions are essentially identical, but framed differently, in terms of gain or loss. If the differential framing has the same effect in this IHL context as it has elsewhere, we might hypothesize that the first group will prefer plan A, whereas the second group will prefer plan B. An outcome like this would be significant not only in the construction of the ‘reasonable military commander’ as a bounded rational actor, but could also contribute to the design of military manuals, for example: the wording of guidelines matters not only for interpretation but for cognition, ‘nudging’ decision-makers to be more or less lenient in targeting approvals. Moreover, an experiment like this could also be used to test whether military commanders with different levels of experience and training have developed rationality that overcomes the ‘trick’ of framing effects, by using different groups from among military commanders and civilians as subjects. To be sure, the analysis itself could suffice as a theoretical application of existing knowledge in cognitive psychology.
Moreover, note that the analysis is non-judgmental in normative terms. It operates within the boundaries of existing legal definitions (of proportionality) and tests the ways in which they can be differentially applied. This, in turn can have normative implications.
I hope these two examples, so different from each other, provide a better understanding of the potential of behavioral analysis of international law. These are only examples, however. I have been asked by many if there are other examples. Of course – I mentioned some in my previous posts, also in reference to the work of other colleagues. But it is not difficult to think of other areas of application. Do international and domestic courts, when faced with competing jurisdiction, consider their authority as an endowment? How could this bear on the positive/negative construction of complementarity in the International Criminal Court? Following on the discussion of proportionality – how do decision-makers consider similar concepts under human rights law – e.g., if an offsetting consideration is phrased as a derogation, a limitation or in terms of proportionality? Does the availability bias skew decisions made with respect to or under international law relating to terrorism and WMD when threats become apparent? Has the positive-listing approach of the General Agreement on Trade in Services played on status quo biases to prevent greater liberalization in the field? I leave to readers and researchers to add to the list.