Trust, Cooperation, and Conflict: Neuropolitics and International Relations

Abstract

Neuroeconomics or the study of brain activity illuminates how choices are made. Kugler and Zak show how oxytocin impacts brain chemistry and levels of trust, with possible implications for decision-making, and, in turn, conflict and cooperation. They offer a challenge to realist theory by questioning its assumptions about a fully Hobbesian world and by incorporating the notion of trust into our thinking and modeling as well. We are misled by the over-emphasis on anarchy at the systemic level to the neglect of sub-systemic variables such as trust and the neuroscience aspects that help shape it. Kugler and Zak argue that willingness to accept international rules that restrict behavior and help avoid conflict depends critically on the level of trust. While they recognize that numerous variables influence cooperation, and that realists certainly do hold out the prospect for cooperation even if that is difficult to achieve, they believe that our understanding of trust is under-developed as is the role of neuroscience in influencing our decisions. Their chapter combines the path-breaking work of Paul Zak in the laboratory with Kugler’s enormous record in international relations.

Appendix 1

The game-theoretic NMT proposed here is based on macro studies on power dynamics and the studies Zak’s laboratory has done on the role of OT in facilitating trust. The model is presented to show how findings from neuroeconomics studies can be included in an otherwise standard optimizing model. The model is a dyadic interaction that captures the essence of trust—that one party must make a decision first and the other party subsequently reacts to this. Our hope is that social scientists will use this model as a foundation to incorporate trust into models of political decisions particularly in the complex arena of deterrence where formal models affect policies that impact the safety of millions.

Let us begin with some notation. The players are identified as decision maker 1 (DM1) and decision maker 2 (DM2). These can be considered countries or individual political actors. Table A1 lists the variables in the model and their definitions. The DMs in this model are self-interested but also value the ongoing relationship between the parties. Both parties have identical continuous, increasing, and strictly concave utility functions U(). Decisions are made with full information by both DMs, and we abstract from stochasticity in this version of the model. For simplicity and consistent with the parity conditions previously identified in power transition dynamics, assume both parties begin with equal resources (M1 = M2 = M > 0) and have a history of interacting with each other, V(R) > 0 and V(R) finite. The results of the model do not change if these two assumptions are relaxed.

We know from Zak’s studies of OT release discussed above that under conditions of low stress, when DM2 receives a transfer from DM1, OT is released and this induces a desire to reciprocate. The degree of reciprocity scales with the size of the transfer. We will call DM2’s reciprocity function

α(S, E, V) : R ⁺ x[0, 1]xR ⁺ → [0, 1]

and assume that it is continuous and increasing in S reflecting status quo level in power dynamics and decreasing in E and V. The shared external environment, E, that both DMs experience affects the reciprocity that DM2 will have toward DM1 by affecting stress levels during the interaction. E is negatively affected by stress that can emanate from an international crisis, imminent elections, a shortage of time to make decisions, or other factors that impact decision-making. As discussed above, high stress reduces OT release and therefore DM2 reciprocity. The implications for the understanding of political exchanges, particularly international exchanges that are currently assumed to be unaffected by reciprocity, are, clearly, fundamental.

Variance of what DM1 sends to DM2, V, captures how trusting DM1 has been in previous interactions with DM2. This permits DM2 to “right old wrongs” by reducing returns R if DM1 has been inconsistent in showing trust. This term R captures the notion that trust takes time to build but is easy to break. The effect of the variance on what is sent causes both DMs to indirectly place value on a continuing relationship; since the game is positive sum, win-win solutions exist if α ₂(S, E, V) is large enough.

The model is solved backwards. DM2’s decision problem is:

Max _{K2 , R} U(K ₂) + α ₂(S, E, V) U(R)

s . t . K ₂ = M + ηM+

DM2 is self-interested because greater resources K ₂ are valued, but at the same time, she/he receives a utility flow U(R) from reciprocating with DM1. This utility occurs even in a one-shot interaction (V = 0) but is reinforced by a continuing relationship.

The necessary and sufficient condition for an optimum is

U ^′(K ₂) = α ₂(S, E, V) U ^′(R)

This determines DM2’s optimal amount kept, ${K}_2^{*}$ and amount returned to DM1, R*. It is straightforward to prove that R* is increasing in α ₂. In other words, for a fixed environment E, the more resources DM2 receives, the greater his/her reciprocity toward DM1, and the larger the amount returned to DM1. Nevertheless, for E sufficiently high, α ₂(S, E, V) = 0 and as a result DM2 uses the self-interested Nash strategy, R* = 0. One can also show that R* is decreasing in the variance, V.

DM1 is also self-interested, gaining utility from total resources she/he keeps K ₁ and from what DM2 returns, R. DM1 understands that for a given environment E, sending an amount S to DM2 will have a neurologic effect on him/her. That is, S will induce OT release that will motivate reciprocity by DM2. DM1 also understands the value of consistency, V, in the trust shown to DM2 and knows that inconsistency will reduce DM2’s reciprocity.

DM1’s decision is to solve

Max _{K1 , S} U(K ₁ + R)

s . t . K ₁ + S = M

R = η = Mδεχ ₂

U ^′(K ₂) = α ₂(S, E, V) U ^′(R)

The optimum is found from the condition

U ^′(K ₁ + R)(ηR)(pt ₂(S)/dS) = 0,

or

ηr/(dS)₂(S)/dS.

This equation reveals that the amount that DM1 sends to DM2, S, depends on the return, η, as well as on how DM2 responds to being shown trust by DM1, K ₂(S). The function K ₂(S) is positive and decreasing in S (equivalently, R(S) is increasing) so that DM1 expects a return from DM2 unless E is high or there is a fractured history of reciprocity (V high).