Two cafés, Tony's and Mary's, face each other across the main square of a small village and serve only espresso, at a cost of $0.10 a cup. There are 200 potential customers, each willing to pay up to $1.50. Half are regulars: 50 always go to Tony's and 50 to Mary's, as long as the price is at most $1.50. The other 100 are shoppers, who go to the cheaper café and split evenly if the prices are equal.
Week 8: Practice
Double Espresso
Question 1
The cafés set prices at the same time and may charge any price. Find Mary's best response to Tony's price. Is there a Nash equilibrium in which each café charges one price for sure?
Reveal the solution
Against Tony's price \(p\), Mary has two sensible replies. She can undercut by a cent and serve all 150 of her regulars and the shoppers, earning about \(150(p-0.10)\). Or she can give up on the shoppers and charge $1.50 to her 50 regulars, earning \(50\times1.40=70\). Undercutting is better when \(150(p-0.10)>70\), that is, when \(p\) is above about $0.57.
So there's no equilibrium with fixed prices. Above $0.57 each café undercuts the other, but once prices reach $0.57, the best reply jumps back to $1.50, and the undercutting starts again. The regulars break the Bertrand result: each café always has a profitable fallback. In practice, such markets show frequent sales and price cycles.
Question 2
The village now allows only two prices, $1.50 and $1.10, and the square closes for renovation in two days, so the cafés have two days left. Each day they set prices at the same time. Tony has put up a binding poster: "If we cut our price tomorrow, we'll refund the difference to everyone who paid more today." Mary has no poster. What are the credible equilibria?
Reveal the solution
Each day's profits, before any refund:
Without posters, this is a prisoner's dilemma with $1.10 dominant, and the last day would be ($1.10, $1.10).
Tony's poster matters on day 2 only if he charged $1.50 on day 1. A cut to $1.10 then costs him $0.40 for each of his day-1 customers. Work out day 2 after each day-1 outcome, then add the two days.
- Both $1.50 on day 1 (100 customers each): on day 2, Mary's $1.10 is dominant. Cutting would cost Tony a refund of 40, so he keeps $1.50 (70 against \(100-40=60\)). Totals: Tony \(140+70=210\), Mary \(140+150=290\).
- Tony $1.50, Mary $1.10 on day 1 (Tony has 50 customers): on day 2 Mary plays $1.10, and Tony cuts too (\(100-20=80\) against 70). Totals: Tony \(70+80=150\), Mary \(150+100=250\).
- Tony $1.10 on day 1: the poster doesn't bind, and day 2 is ($1.10, $1.10). Totals: Tony 250 and Mary 170 if Mary charged $1.50, or 200 each if both charged $1.10.
$1.10 is dominant for Tony on day 1, and Mary's best response is $1.10. In the credible equilibrium, both charge $1.10 on both days and each earns 200. One poster isn't enough.
Question 3
Can Mary's choice of whether to put up the same poster support both cafés charging $1.50 on both days?
Reveal the solution
Yes. With both posters, after ($1.50, $1.50) on day 1, cutting on day 2 costs each café a refund of 40, so $1.50 is dominant for both on day 2 (140 against 110, and 70 against 60). The other day-1 outcomes work as before.
There are now two equilibria: ($1.50, $1.50) on both days with 280 each, and ($1.10, $1.10) with 200 each. Mary puts up the poster, and both cafés charge $1.50 on both days.
The poster is a commitment that looks like a gift to customers but works against them: it makes cutting the price expensive, so neither café cuts. Most-favored-customer clauses and price-matching guarantees can have the same effect. It's the logic of Hats, Alcoa's capacity and Kiwi's small entry: a visible, costly promise that changes what you'll want to do later.