Goals of the course: The course will be a mix of modern IO theory and empirical observations about the structure of firms and markets, and behavior by firms that are not perfectly competitive. The majority of recent developments in IO have been theoretical in nature, arising from applications of game theory to stylized environments. Relative to labor economics, empirical work in IO is weak, and research topics in empirical IO abound. The course will stress a basic understanding of the theory, with attention paid to the defects of the theory and opportunities for useful empirical research.
1. What is industrial organization?
2. Organization of the firm.
Whether a mix of activities is carried out by one firm, or by several distinct firms (tied by contracts, perhaps) affects the resolution of disputes, the distribution of profits not contracted on (residual claims) and the decision-making if events not contracted on arise.
After a review of the standard monopoly model, we will consider several topics in price discrimination. Price discrimination may be based on observable characteristics (e.g. mailing coupons to certain addresses) or operate through self-selection (e.g. quantity discounts and newspaper coupons available to all), or both. Spatial price discrimination will be analyzed as well. An issue related to price discrimination, monopoly bundling (Shoe: buy one, get one free) will be analyzed. The welfare impact of price discrimination will be considered. Price discrimination is illegal in the United States, although the law admits many exceptions. The design and regulation of monopolies will be considered in environments where the regulator knows less about the costs of the firms than does the monopoly.
4. Strategic Behavior in Oligopoly
Unlike the competitive or monopoly models, oligopoly models are based on the understanding that firms recognize and react to the effect that their decisions have on other firms. We will start with the simplest models, n firms competing to sell a homogenous good to a large group of consumers. Solutions to the static problem will be developed (under quantity and price competition), and then the use of trigger strategies and meet or beat pricing to support high prices considered.
5. Merger Analysis and Antitrust
Why do firms merge? What are the laws and regulations governing mergers? Only a few of the usual reasons stated for mergers have been formally modelled, but most the reasons are interesting, and some even seem plausible. The laws and regulations governing mergers (and, indeed most economic activity) often seem to have little connection to a sensible economic analysis, although economic analysis is used extensively to come up with these laws and regulations.
A risk neutral insurer offers insurance to an agent, who knows his own riskiness, and the insurer does not. This is the classic asymmetric information model, simple to solve, and illustrates many of the effects of asymmetric information: an full-information efficient contract only for one possible type of agent, inefficiently little insurance for others.
7. Conspiracy, Cartels and Tacit Collusion
One of the triumphs of game theoretic modelling of firm behavior was understanding how competing firms might profitably not compete with each other, by the use of implicit threats and tacit collusion. This understanding overturned a widely held but erroneous belief that cartels must eventually fail (such claims were made, even about OPEC, which, of course, did in fact fail as a successful cartel). However, the power of the theory, even to operate in environments with limited observability of rival's actions and incomplete ability to punish defectors, left an equally large mystery: why don't cartels always succeed, and why do they occasionally fall apart?
In addition to the theoretical analysis of collusion, we will consider the laws governing conspiracy and "bid-rigging". There is a second empirical mystery about conspiracy, which is that it typically occurs in industries with easy entry (movie theaters, road building, moving and storage, etc.), where presumably entry undermines the collusive profits.
There are two major theories of advertising. In the burning money theory, a firm tries to signal that it has a good product by advertising more than the seller of a bad product could profitably do, i.e. signalling that it expects repeat business. Such advertisements need not have any content: they are just supposed to be expensive. Price advertising may yield imperfectly informed consumers, and in many such models, the equilibrium involves randomized prices, which gives a model of sales.
9. Entry and Exit
A classic issue in IO is the deterrence of entry and the encouragement of exit by pricing decisions. This will be analyzed using the game theoretic tools developed for pricing behavior. The modern formulation of the entry deterrence theory involves an incumbent who knows more about the cost or demand than a potential entrant, and the entrant uses the incumbent's price as a signal of this cost or demand, creating an ability for the incumbent to deter entry by changing his price. However, the entrant understands this ability as well. The analysis of this question originated with court decisions on predatory pricing: could firms profitably drive their rivals out of business, and then enjoy monopoly profits? If consumers were hurt by this behavior, it would be illegal under the Clayton Act.
10. Product Differentiation and Patents
Offering a different product than other firms creates a degree of market or monopoly power. Natural questions concern how firms behave when they can differentiate their products. Do they produce too many or two few different products? Can product differentiation act as a barrier to entry (spatial preemption)? A patent creates a temporary monopoly, which creates an incentive for firms to invest in research to win a patent race. Will firms invest too much or too little in these races?
11. Auctions and Bidding
The issue of how prices are set in imperfect competition is a fundamental issue for understanding how markets work. Auctions provide one important model of price determination. A significant aspect of behavior in auctions concerns how a bidder thinks about his rivals' behavior, because a given bidder typically does not know the rivals' values for the good. One example of this strategic interaction is the winner's curse, which, simply stated, means that the bidder who most overvalues the object for sale will win the bidding. This means that bidders must bid well below their estimate of the value, for the fact that other bidders will not pay this price is evidence that the object has low value. How does the expected price varies with the number of bidders? What are the effects of providing information about the value of the good on bidding? What are the effects of revealing the number of bidders on the bidders? What are the effects of setting an entry fee or a reserve price?
12. Agency Theory
Agency theory, or principal-agent models, are an important tool for understanding contracting. There are two main types of agency model. In the first, a risk-neutral principal (e.g. insurance company) contracts with a risk-averse agent. The contract must balance risk-sharing (the principal is more efficient at absorbing risk) with the agent's incentive to exert effort (a contract in which the agent's payoff is constant puts all the risk on the principal, but gives the agent no incentive to exert effort). In the other class of models, both the principal and the agent are risk-neutral, but the agent knows more about his cost of effort than the principal. As a consequence, the agent earns informational rents from the contract (low cost agents can act like high cost agents, thereby being paid more than their cost), and the principal has an incentive to distort the agent's effort to reduce the informational rents, and full information efficiency is not obtained. Several results emerge from this analysis. More able agents work harder, and only the best possible agent has a full-information efficient contract. More able agents have higher commissions and lower salaries than less able agents. Agency theory is also the theory of optimal taxation.
13. Bilateral Bargaining and The Market for Lemons
A buyer and seller each know their value of an object that the seller holds, but do not know the other's value. It turns out that there is no mechanism for achieving efficient trade; that is, it is not possible for these two to trade whenever trade is socially optimal, because each has an incentive to distort the mechanism to improve the terms of trade. In a dynamic context, this leads to delay in trade. The general thrust of this topic is to understand how markets work when information is not available to all.