Ståhl's bargaining model (Q1175962)

The paper deals with non-cooperative bargaining games with two players and a finite horizon as introduced by \textit{I. Ståhl} [``Bargaining theory'', Stockholm School of Economics, Stockholm (1972)] (i.e., bargaining must end after one week). The disagreement outcome is (0,0) and the feasible set is given by \(S=\{(u_ 1,u_ 2): u_ 1\geq 0, u_ 2\geq 0, u_ 2\leq F(u_ 1)]\), where \(F\) is a strictly concave and decreasing function. Let the bargaining rounds within one week be numbered by \(1,2,\dots,T\), \(T<\infty\), and denote by \(e^{-r}\), \(r\geq 0\), the discount factor at the end of the week. For each \(t\leq T\), if play has continued to round \(t\), a fair coin is tossed. If heads come up, player 1 suggests an outcome \((x^ t,F(x^ t))\in S\). If player 2 accepts, the game terminates and the discounted payoffs are given by \((e^{-rt/T}x^ t, e^{-rt/T}F(x^ t))\). If player 2 rejects in round \(t=T\), payoffs are the disagreement outcome (0,0). If player 2 rejects in round \(t<T\), the game moves on to round \(t+1\). If tails come up in round \(t\), the roles are reversed and player 2 suggests \((y^ t,F(y^ t))\in S\). The main result states that, if the time interval between successive offers approaches zero, the subgame perfect equilibrium payoffs approximate the so-called Raiffa bargaining solution. In other words, Ståhl's finite horizon bargaining model with random offers provides a non-cooperative foundation for the well-known Raiffa bargaining solution.