The Rubinstein bargaining model captures a simple but powerful situation: two players take turns proposing how to split a surplus (say, a dollar), and the other side can accept or reject. If the offer is accepted, the game ends and each player gets their agreed share. If it's rejected, the game moves to the next round, roles flip, and the other player now proposes.

A few key assumptions hold the model together:

There are exactly two players with perfect information about each other's preferences.
Both players are rational and want to maximize their own payoff.
The game has a potentially infinite horizon, meaning there's no fixed deadline. Rounds can continue indefinitely until someone accepts.
Delay is costly. Each round that passes without agreement reduces the value of the surplus to both players (this is where discount factors come in).

That infinite horizon matters. If the game had a known final round, you could solve it with simple backward induction from that endpoint. The elegance of Rubinstein's result is that even with infinite rounds, there's a unique equilibrium.

Discount Factors and Impatience

Discount factors are the engine of the model. They capture how much each player cares about getting a deal now versus later.

Player 1's discount factor is $\delta_1$ and Player 2's is $\delta_2$ , where each $\delta$ falls between 0 and 1.
A high $\delta$ (close to 1) means the player is patient. Waiting one more round barely reduces the value of the surplus for them.
A low $\delta$ (close to 0) means the player is impatient. Every round of delay is expensive.

Think of it this way: if you value a dollar today at a dollar, but a dollar next round at only $\delta$ of a dollar, then $\delta = 0.9$ means waiting costs you 10% per round, while $\delta = 0.5$ means waiting costs you 50%.

In labor negotiations, for instance, the discount factor captures the cost of a strike. A union with a large strike fund can hold out longer (higher $\delta$ ), while an employer facing urgent production deadlines loses more each day without a deal (lower $\delta$ ).

Subgame Perfect Equilibrium in the Rubinstein Model

Subgame Perfect Equilibrium Concept

A subgame perfect equilibrium (SPE) requires that each player's strategy is optimal at every decision point in the game, not just at the start. This rules out non-credible threats. For example, a responder can't credibly threaten to reject every offer forever, because at some point accepting would make them better off.

In the Rubinstein SPE, the proposer in each round offers the responder exactly enough to make them indifferent between accepting now and rejecting (which would let the responder propose next round). The responder can't do better by waiting, so they accept immediately.

Model Overview and Assumptions, Negotiation - Praxis Framework

Deriving the SPE Division

Here's how to find the equilibrium shares, step by step:

Set up the indifference conditions. When Player 1 proposes, Player 2 must be offered at least what Player 2 would get by rejecting and proposing next round (discounted by $\delta_2$ ). The same logic applies in reverse when Player 2 proposes.
With equal discount factors ( $\delta_1 = \delta_2 = \delta$ ): The SPE gives the first-round proposer a share of $\frac{1}{1+\delta}$ and the responder gets $\frac{\delta}{1+\delta}$ . The proposer always gets more, but only slightly more when $\delta$ is close to 1.
With different discount factors: The SPE shares for Player 1 (proposing first) and Player 2 are:

$x_1 = \frac{1 - \delta_2}{1 - \delta_1 \delta_2}, \quad x_2 = \frac{\delta_2(1 - \delta_1)}{1 - \delta_1 \delta_2}$

Agreement is immediate. In equilibrium, Player 2 accepts the very first offer. There's no delay, even though the game could go on forever.

Two critical results emerge from these formulas:

First-mover advantage: The proposer captures a larger share. But this advantage shrinks as both players become more patient ( $\delta \to 1$ ).
Patience pays: The more patient player (higher $\delta$ ) gets a larger share. If $\delta_1 > \delta_2$ , Player 1 does better not just from proposing first, but from being more willing to wait.

The SPE is unique, which is one of the model's most celebrated properties. Unlike many bargaining setups where multiple equilibria exist, Rubinstein's alternating-offers structure pins down a single outcome.

Example: Numerical Illustration

Suppose $\delta_1 = 0.9$ and $\delta_2 = 0.7$ , and the surplus is 1.

$x_1 = \frac{1 - 0.7}{1 - (0.9)(0.7)} = \frac{0.3}{1 - 0.63} = \frac{0.3}{0.37} \approx 0.81$

$x_2 = 1 - 0.81 = 0.19$

Player 1, being more patient, captures roughly 81% of the surplus. Player 2's impatience costs them dearly.

Patience and Bargaining Outcomes

Impact of Patience on Bargaining Outcomes

The core insight of the Rubinstein model is that patience is power. The player who can afford to wait longer extracts a larger share, because the other side knows that rejecting an offer just leads to a worse deal next round (after discounting).

As both discount factors approach 1 (both players become very patient), the first-mover advantage vanishes and the split approaches 50-50. This makes intuitive sense: if neither side cares much about delay, there's no leverage from being the proposer.

Conversely, if one player is extremely impatient ( $\delta$ close to 0), the patient player captures nearly the entire surplus. The impatient player would rather accept almost anything now than wait.

Model Overview and Assumptions, Rational Decision Making vs. Other Types of Decision Making | Principles of Management

Outside Options and Bargaining Power

The basic Rubinstein model assumes players have no alternative to continued bargaining. But real negotiations almost always involve outside options, a payoff a player can guarantee by walking away from the table.

Outside options affect the equilibrium in a specific way:

If a player's outside option is less than what they'd get in the Rubinstein SPE, it doesn't change the outcome. The equilibrium already gives them more than walking away would.
If a player's outside option exceeds their Rubinstein share, it becomes a binding constraint. The other player must offer at least that much to prevent a walkout, shifting the division.

This is sometimes called the outside option principle: outside options matter only when they're actually attractive enough to be used.

In salary negotiations, for example, an employee with a strong competing job offer has a credible outside option. The employer must beat that offer to keep the employee at the table, which shifts bargaining power toward the employee.

Applying the Rubinstein Model to Real-World Situations

Labor Negotiations

The Rubinstein framework maps naturally onto union-employer disputes. The union proposes a wage, management can accept or counter-offer, and each round of failed negotiation means a costly strike or lockout.

A union with a large strike fund has a higher discount factor: it can sustain a work stoppage longer, which strengthens its position.
An employer facing time-sensitive contracts or perishable inventory has a lower discount factor: delay is disproportionately costly, weakening its bargaining position.
The model predicts that strikes should be rare in equilibrium (agreement is immediate), which matches the empirical observation that most labor negotiations settle without a strike. When strikes do occur, it often points to incomplete information, where one side misjudges the other's patience.

Mergers and Acquisitions

In M&A negotiations, the acquiring firm and the target company bargain over the acquisition price.

The cost of prolonged negotiations (management distraction, market uncertainty, regulatory timelines) determines each side's discount factor.
A target company with a competing bidder (a "white knight") has a valuable outside option, forcing the acquirer to offer a higher price.
An acquirer that has already announced its intent publicly may face reputational costs from walking away, effectively lowering its outside option and weakening its position.

Extensions of the Rubinstein Model

The basic model makes strong assumptions (perfect information, two players, risk neutrality) that rarely hold perfectly in practice. Several extensions address this:

Incomplete information: If players don't know each other's discount factors, bargaining can involve delay as a signaling mechanism. Patient players may reject early offers to reveal their type, which explains why real negotiations sometimes drag on.
Multiple players: Coalition bargaining (e.g., in parliamentary politics) requires adapting the two-player framework. The key challenge is that any two players can potentially form a coalition and exclude the third.
Risk aversion: Risk-averse players behave similarly to impatient ones. They'd rather lock in a certain payoff now than gamble on getting more later, which weakens their bargaining position.

These extensions bring the model closer to real-world complexity while preserving its central insight: the distribution of patience and alternatives determines who gets what.

2,589 studying →