ECO100 - 15 - F2026
Overview
Matthew Baker uses the Hotelling and Bertrand competition models to show how firms’ best responses can produce clustered locations, competitive prices, or limit pricing that deters entry. He then applies backward induction to sequential games and compares game-theory predictions with laboratory behavior in centipede, ultimatum, and investment games, highlighting how individual incentives can conflict with outcomes that benefit everyone.
Key takeaways
- Hotelling competition predicts that two identical vendors on a uniformly populated beach cluster at the midpoint, even though quarter-point locations reduce the maximum customer walk to one-quarter of the beach.
- Bertrand competition with identical goods and costs drives firms to price at marginal cost because either firm can capture the market by undercutting its rival.
- A low-cost firm can use limit pricing by charging just below a rival’s costs, preserving its own margin while making entry unprofitable for the higher-cost rival.
- Backward induction makes threats credible only when carrying them out is rational at the point of decision; this is why the entrant enters when the incumbent would rather accommodate than start a price war.
- The centipede game’s backward-induction outcome is an immediate take, even though passing can grow the available payoff to 5 for Player 1 and 3 for Player 2.
- Ultimatum and investment experiments often show more fairness and trust than basic self-interest predictions: unfair splits are rejected, and participants may invest or return money without enforcement.
Chapters
- A Nash equilibrium occurs when each player is choosing a best response to the other, so neither wants to change actions alone.
- Game theory models players, strategies, and payoffs, and can analyze situations beyond payoff matrices.
- The November 11 exam covers monopoly and some game theory from after the first exam; review materials are planned for the following week.
- Harold Hotelling’s spatial competition model places two identical hot-dog carts on a one-unit beach with consumers distributed uniformly.
- Because customers buy from the nearest cart and prices and products are identical, each cart’s location determines its share of beachgoers.
- A cart can gain customers by moving closer to its competitor and then positioning just beyond it, prompting the other cart to respond in turn.
- The location game settles with both carts side by side at the beach’s midpoint, where neither cart can gain by moving alone.
- The model helps explain why gas stations and fast-food restaurants often cluster at the same highway exit.
- The midpoint outcome is less convenient for beachgoers than locations at one-quarter and three-quarters of the beach, which limit each customer’s walk to one-quarter of its length.
- In the Bertrand model, two firms sell identical products and customers buy from whichever charges the lower price.
- When one firm charges more, its rival can undercut it slightly and capture the market, giving the first firm an incentive to undercut in response.
- With identical costs, this price-cutting process reaches a Nash equilibrium at marginal cost, leaving both firms with zero economic profit.
- When Firm A has lower production costs than Firm B, A can price just below B’s costs while staying above its own.
- At that price, Firm A serves the whole market and Firm B cannot profitably match the price, producing the limit-pricing outcome.
- Baker connects this logic to Microsoft’s defense in a government monopolization case: Microsoft argued that low margins discouraged entry and constrained its pricing.
- Dynamic games differ from simultaneous games because players act sequentially and later choices depend on earlier actions.
- Game trees, also called extensive-form representations, show the possible actions and resulting payoffs at each decision point.
- Backward induction identifies rational choices from the end of the tree toward the beginning, helping distinguish credible threats from threats a player would not carry out.
- In the entrant-monopolist example, an incumbent prefers cooperation to a costly price war after entry, so the entrant anticipates cooperation and enters.
- In Baker’s parent-child example, the parent dislikes punishing the child, making the threat not to punish after disobedience non-credible.
- In the startup example, the incumbent innovates whether the startup chooses a traditional method or innovates, so backward induction predicts that the startup innovates and both firms innovate.
- In the centipede game, players can take the current payoff or pass, allowing the available payoffs to grow over successive turns.
- Backward induction predicts that Player 2 takes at the final decision, earning 4 rather than passing for 3; anticipating this, Player 1 takes at the preceding turn.
- Reasoning back through every turn predicts that Player 1 takes immediately, even though continued passing could produce a final payoff of 5 for Player 1 and 3 for Player 2.
- In the arms-race example involving the United States and the USSR, arming is a dominant strategy for each, although mutual disarmament would give both a payoff of 3.
- The prisoner’s dilemma illustrates how individually rational choices can leave everyone worse off than cooperation would.
- Laboratory research also examines dictator, trust, and endowment-effect games; in a coffee-mug experiment, people often demand more to sell a mug they own than they initially said they would pay for it.
- In the $10 ultimatum game, the proposer offers the responder an amount X and keeps $10 minus X; rejection gives both players zero.
- Under the basic rational-payoff prediction, a responder accepts any positive offer, so the proposer offers the smallest possible positive amount.
- Experiments commonly depart from the predicted 99-to-1 split: responders reject strongly unequal offers, and proposals are more often around 50/50 or 60/40.
- In the investment game, an investor gives amount X to an entrepreneur, chance may triple the funds, and the entrepreneur chooses how much R to return.
- Without enforcement or contracts, backward induction predicts a return of zero and therefore an initial investment of zero.
- People in experiments often invest and return some money anyway; Baker closes by contrasting these cooperative behaviors with strategic models and previews externalities for the next class.
Transcript excerpt
0:00 So, my deepest apologies there for the delay. But, it happens once a semester, so I think we we've made it this far, which is uh uh pretty good. Uh anyways, I'll try to just speed things up a little bit. I think we're still going to be okay time-wise. Uh as my plan today is to just say some more stuff about uh game theory. Uh anyways, so today is the 2nd, right, uh of November, and we do have our next exam scheduled the 11th. So, this coming Wednesday, I will give you the review uh materials. I'll post some um questions, some online questions as well, and we can go over all of that a week from today, and then you'll know what'll happen. It'll be about the same. Uh and hopefully, when you get the practice materials, you'll see uh about what the form and content of
0:49 the exam will be. Suffice it to say, it is going to be all about monopoly uh and a little bit of game theory, and whatever else we can say on uh on the 4th. Uh not nothing from the first part of the class, right? Just everything since uh our first exam until the present. Uh but, we're going to we're going to move ahead here quickly, so we can talk about all of that with the review. So, we talked last time a bit about this concept of Nash equilibrium. And essentially, what we do with game theory is we think about how rational players will interact strategically using some particular set of tools, right? First of all, there's this whole notion of a game, right? A thing with payoffs and
1:36 strategies. The payoffs depend upon whatever actions, right? Actions together comprise potential strategies that players may take.
This is the start of the transcript. Save this video free for the whole transcript, searchable and time-coded.
In this course
Related notes
Summary, takeaways, and chapters were generated by AI from the video's transcript and may contain errors. The video belongs to its creator, Matthew Baker.