Skip to content
Tutorials Center
My learning

Topics

Lesson 22 of 37 ยท Sequential Games

Stackelberg Competition and Commitment

Solve the Stackelberg leader-follower game with backward induction, compare it with Cournot, and see how commitment and entry deterrence create a first-mover advantage.

In some markets one firm chooses its output first and the rival observes it before choosing its own. That is the Stackelberg model. It is the most famous example of a sequential game with continuous actions, and it shows something deep: being able to commit to an action can make you better off. The leader's advantage does not come from being smarter. It comes from moving first in a way the follower must take as given.

The model

Two firms produce the same good. The market price is P = a - (q1 + q2), where q1 and q2 are the quantities. Each firm has a constant cost c per unit, and a > c. Profit of firm i is (P - c) times qi.

Simultaneous (Cournot) version. Each firm picks its quantity without seeing the other. Setting the derivative of profit to zero gives the best-reply functions q1 = (a - c - q2) / 2 and q2 = (a - c - q1) / 2. The Nash equilibrium is q1 = q2 = (a - c) / 3.

Sequential (Stackelberg) version. Firm 1 (the leader) chooses q1. Firm 2 (the follower) sees q1 and plays its best reply. By backward induction, the leader plugs the follower's reaction into its own profit and maximizes that.

Stackelberg in three steps

Leader moves Picks quantity q1
Follower reacts Best reply to q1
Leader foresees it Maximizes with the reaction built in

Solve the follower first, then the leader.

1 / 2

Solving it symbolically

Sympy can derive the solution exactly. Here it first finds the follower's best reply, then substitutes it into the leader's profit.

Python Stackelberg equilibrium with sympy

Press Run (or Ctrl + Enter) to see the result here.

Program input ยท what you'd type while the program runs

Type the answers before you press Run, one per line. Each time the program reads a line (with input()), it gets the next line from this box.

The leader produces (a - c) / 2, the follower (a - c) / 4. With a = 100 and c = 10: leader 45, follower 22.5, price 32.5, leader profit 1012.5 and follower profit 506.25. In Cournot, each produces 30 with profit 900.

You do not need calculus. A grid search over the leader's quantity, with the follower's best reply computed on a grid too, gives the same answer. This approach also works when the demand curve is not linear.

Python Grid search for the Stackelberg outcome

Press Run (or Ctrl + Enter) to see the result here.

Program input ยท what you'd type while the program runs

Type the answers before you press Run, one per line. Each time the program reads a line (with input()), it gets the next line from this box.

Why moving first helps here

At the Stackelberg outcome the leader earns more than in Cournot (1012.5 compared with 900) and the follower earns less (506.25 compared with 900). The leader pushes out more, and the follower, seeing it, reduces its own output because a larger rival output lowers the price. The key is commitment: if the leader could change q1 after seeing q2, the outcome would revert to Cournot. The advantage exists only because the leader's choice is irreversible and the follower knows it.

Notice that more information is not always better. Here the follower knows strictly more (it sees q1) and still earns less than in Cournot. Information can hurt when it changes what the other party chooses.

A caution: moving first is not always an advantage. In games like rock-paper-scissors, going second wins. In Stackelberg with strategic substitutes (a higher rival output lowers your best reply), a first-mover advantage exists. With some other payoff structures (strategic complements, for instance price competition with differentiated goods), leading can be a disadvantage.

Entry deterrence by commitment

Return to the entry game from Extensive-Form Games: the entrant enters, and the incumbent accommodates, because fighting is not credible. Could the incumbent change that? Yes, by changing its own payoffs before the game, for example by building extra capacity, signing a contract that forces low prices, or hiring a manager rewarded for market share. If fighting becomes the best action once entry occurs, the threat becomes credible.

Python Commitment turns a threat credible

Press Run (or Ctrl + Enter) to see the result here.

Program input ยท what you'd type while the program runs

Type the answers before you press Run, one per line. Each time the program reads a line (with input()), it gets the next line from this box.

In the second game, capacity costs the incumbent 1 in every outcome, but it raises the payoff of fighting from -1 to 4, so fighting now beats accommodating (3 against 1). The entrant, anticipating the fight, stays out. The incumbent ends with 3 instead of 2. The commitment made the incumbent's own options worse in some branches (it cannot easily accommodate) and still made it better off overall. That is the paradox of commitment: reducing your options can raise your payoff, which is why people burn bridges and sign binding contracts.

Real-world patterns

  • A firm announces output or capacity before the rival decides (the leader-follower pattern of some dominant-firm industries).
  • Governments announcing a policy rule so that private players adjust to it.
  • A negotiator who publicly commits to a position and removes the option to back down.

All of these rely on the follower believing the move is irreversible.

Check your understanding

Quick check ยท 4 questions

of 4 answered You scored / 4 ยท perfect!

  1. 1. In the linear Stackelberg model, how does the leader's quantity compare with its Cournot quantity?

    With a = 100 and c = 10, the leader produces 45 against 30 in Cournot, because it can commit and the follower will respond by cutting output.
  2. 2. How do we solve a Stackelberg game?

    This is backward induction: the leader foresees the follower's reaction.
  3. 3. What makes the leader's advantage possible?

    If the leader could revise after seeing the follower, the outcome would return to Cournot.
  4. 4. A capacity investment reduces an incumbent's payoff from accommodating and raises its payoff from fighting. Why can it help?

    Once fighting is optimal after entry, the entrant expects a fight and stays out, so the commitment pays.

Summary

  • Stackelberg: the leader commits to a quantity, the follower observes and replies, and backward induction gives the solution.
  • With linear demand, the leader produces (a - c) / 2 and the follower (a - c) / 4. The leader earns more than in Cournot, the follower less.
  • The first-mover advantage comes from commitment, not from information.
  • Changing your own payoffs or removing options in advance can make threats credible and deter entry.
  • Moving first is not always an advantage; it depends on the strategic structure.

Next: The Repeated Prisoner's Dilemma