Imagine two people trying to split a pie, but there’s a twist: the longer they take to agree, the smaller the pie gets. Every moment of delay costs them both, creating pressure to reach a deal quickly. Yet each person wants the largest slice possible. How do they navigate this tension between patience and greed? This is the puzzle at the heart of sequential bargaining, and the Rubinstein bargaining model provides an elegant answer to how rational negotiators behave when time itself becomes a factor in their decisions.
Table of Contents
What makes sequential bargaining different
Most of us have experienced bargaining in some form, whether negotiating a salary, haggling at a market, or deciding with a friend where to eat dinner. But sequential bargaining has a unique structure that distinguishes it from one-shot negotiations. In this framework, two parties take turns making offers to each other about how to divide a resource. If one person rejects an offer, they get to make a counteroffer in the next period. This alternating pattern continues until someone accepts a proposal.
The key insight is that waiting has a cost. Because people are impatient, they value rewards received today more highly than identical rewards received tomorrow. Economists capture this through discount factors, which measure how much less valuable a future payoff is compared to an immediate one. A discount factor close to one means someone is very patient, while a factor closer to zero signals strong impatience. This impatience creates an incentive to reach agreement quickly rather than holding out indefinitely for better terms.
Understanding the Rubinstein model through a simple example
The Rubinstein bargaining model, introduced by economist Ariel Rubinstein in 1982, analyzes exactly this scenario. Consider two players negotiating over how to divide a pie of size one. Player 1 makes the first offer, proposing to keep a certain share for herself and offering the rest to Player 2. If Player 2 accepts, the game ends and they split the pie accordingly. If Player 2 rejects the offer, he gets to make a counteroffer in the next period.
Here’s where impatience matters crucially. Both players discount future payoffs by a factor delta, meaning a pie worth one unit today is only worth delta units tomorrow. This discounting captures the idea that delay is costly, perhaps because of uncertainty, opportunity costs, or simple time preference. The critical question becomes: what offers will the players make, and when will they agree?
How to solve a finite bargaining game
To understand the logic of sequential bargaining, let’s start with a simplified version where there are only two periods. This finite-horizon game can be solved using backward induction, a technique where we reason backwards from the end of the game to determine what rational players should do at each stage.
In the final period, Player 1 has all the bargaining power because this is the last chance to reach a deal. If Player 2 rejects the offer, both players get nothing. Therefore, Player 1 can offer Player 2 the smallest possible amount (essentially zero), and Player 2 will accept it rather than walk away empty-handed. Player 1 captures virtually the entire pie in this final round.
Now step back to the first period. Player 2 knows that if he rejects Player 1’s initial offer, he’ll get to make an offer in period two, but Player 1 can then extract almost everything in the final period. Since payoffs in period two are discounted by the factor delta, Player 2 anticipates receiving close to zero in discounted terms. This means Player 1 doesn’t need to offer much in the first period to make Player 2 willing to accept immediately.
The equilibrium offer that emerges makes Player 2 exactly indifferent between accepting now and rejecting to make a counteroffer later. Player 1 proposes keeping approximately 1/(1+delta) of the pie and offering delta/(1+delta) to Player 2. Player 2 accepts this offer immediately because waiting would give him no advantage. The first mover gains a significant edge in this setup.
Why backward induction is powerful
The technique of backward induction works because it identifies strategies that form a subgame perfect equilibrium. This refinement of Nash equilibrium requires that players’ choices must be optimal not just at the start of the game, but at every possible decision point. In other words, no player should have an incentive to deviate from their strategy at any stage, even after unexpected moves by the opponent.
Backward induction ensures that threats and promises are credible. If a player threatens to reject an offer, that threat must actually be in their best interest when the time comes to carry it out. This eliminates strategies based on empty threats that would hurt the threatener if actually executed. By working backwards from certain endgames, we can determine which strategies are truly credible throughout the entire negotiation.
The surprising result with infinite periods
The real elegance of the Rubinstein model emerges when we extend the game to an infinite horizon, where players can keep making alternating offers indefinitely. At first glance, this seems impossibly complex to solve. With no final period to anchor the backward induction reasoning, how can we determine what rational players should do?
Ariel Rubinstein showed that even with infinite periods, the game has a unique subgame perfect equilibrium. The solution relies on a clever observation about stationarity. Because the structure of the game is identical in every period, the equilibrium must have a particular self-consistency property. If we know what split Player 2 would offer in period two, we can work out what Player 1 should offer in period one to make Player 2 just willing to accept immediately.
The equilibrium formula is remarkably clean. Player 1 receives 1/(1+delta) of the pie, while Player 2 gets delta/(1+delta). Notice how this depends entirely on the discount factor. When delta is close to one, meaning players are very patient, the split approaches fifty-fifty. When delta is smaller, indicating more impatience, Player 1’s first-mover advantage grows larger. The more costly delay is, the more valuable it becomes to propose first and avoid waiting.
How patience affects bargaining power
An important extension allows the two players to have different discount factors. Perhaps one negotiator faces tighter deadlines or has better outside options. In this case, the more patient player captures a larger share of the surplus. If Player 1 has discount factor delta₁ and Player 2 has discount factor delta₂, the equilibrium gives Player 1 a share of 1/(1+delta₂).
This result has intuitive appeal. The player who can afford to wait longer has more bargaining power because they’re less desperate to reach immediate agreement. In labor negotiations, for example, a union with substantial strike funds might be more patient than a company facing time-sensitive production deadlines, giving the union leverage. In international diplomacy, a country with stable domestic support might negotiate more effectively than one facing political pressures for quick results.
Why this model matters beyond theory
The Rubinstein model has influenced how economists think about many real-world situations. Labor negotiations, international trade agreements, merger discussions, and even everyday commercial transactions all involve elements of sequential bargaining. The model’s predictions about first-mover advantages, the importance of patience, and the role of time costs in negotiations provide practical insights for anyone engaged in strategic interactions.
One particularly important implication is that even though the game could theoretically continue forever, rational players reach agreement immediately. There’s no delay in equilibrium because any postponement makes both parties worse off through discounting. This “no-delay” result suggests that when we observe prolonged negotiations in reality, something beyond pure rationality must be at play, perhaps incomplete information about the other party’s patience or preferences, or behavioral factors like spite or fairness concerns.
The model also connects non-cooperative bargaining theory with cooperative approaches like the Nash bargaining solution. As players become infinitely patient and discounting disappears, the Rubinstein outcome converges toward an equal split, which matches predictions from axiomatic bargaining theory. This convergence shows that different theoretical frameworks can sometimes reach similar conclusions from very different starting points.
What do you think? Have you ever been in a negotiation where timing and patience played a crucial role in the outcome? How might understanding the formal structure of sequential bargaining change the way you approach real-world negotiations?
Leave a Reply