Y 0 Tax{ Compensated Choice qa. q8 Introduction to Economic Analysis by R. Preston McAfee J. Stanley Johnson Professor of Business, Economics & Management California Institute of Technology 2 .02 's Unstable Equilibrium7 Stable Equilibrium McAfee: Introduction to Economic Analysis, http://www.introecon.com, July 24, 2006 Dedication to this edition: For Sophie. Perhaps by the time she goes to university, we'll have won the war against the publishers. Disclaimer: This is the third draft. Please point out typos, errors or poor exposition, preferably by email to intro dmcafee.cc. Your assistance matters. In preparing this manuscript, I have received assistance from many people, including Michael Bernstein, Steve Bisset, Grant Chang-Chien, Lauren Feiler, Alex Fogel, Ben Golub, George Hines, Richard Jones, Jorge Martinez, Joshua Moses, Dr. John Ryan, and Wei Eileen Xie. I am especially indebted to Anthony B. Williams for a careful, detailed reading of the manuscript yielding hundreds of improvements. McAfee: Introduction to Economic Analysis, http://wwwintroecon.com, July 24, 2006 jj This work is licensed under the Creative Commons Attribution- NonCommercial-ShareAlike License. To view a copy of this license, visit http://creativecommons.org/licenses/by-nc-sa/2.O/- or send a letter to Creative Commons, 559 Nathan Abbott Way, Stanford, California 94305, USA. Please email changes to intro@mcafee.cc. Introduction to Economic Analysis Version 2.0 by R. Preston McAfee J. Stanley Johnson Professor of Business, Economics & Management California Institute of Technology Begun: June 24, 2004 This Draft: July 24, 2006 This book presents introductory economics ("principles") material using standard mathematical tools, including calculus. It is designed for a relatively sophisticated undergraduate who has not taken a basic university course in economics. It also contains the standard intermediate microeconomics material and some material that ought to be standard but is not. The book can easily serve as an intermediate microeconomics text. The focus of this book is on the conceptual tools and not on fluff. Most microeconomics texts are mostly fluff and the fluff market is exceedingly over- served by $1oo+ texts. In contrast, this book reflects the approach actually adopted by the majority of economists for understanding economic activity. There are lots of models and equations and no pictures of economists. McAfee: Introduction to Economic Analysis, http://www.introecon.com, July 24, 2006 jjj Table of Contents 1 WHAT IS ECONOMICS? ..............................1-1 1.1.1 Normative and Positive Theories ..................1-2 1.1.2 Opportunity Cost ..........................................1-3 1.1.3 Economic Reasoning and Analysis................1-5 2 SUPPLY AND DEMAND...............................2-8 2.1 Supply and Demand.................................2-8 2.1.1 Demand and Consumer Surplus ...................2-8 2.1.2 Supply .........................................................2-13 2.2 The M arket...............................................2-18 2.2.1 Market Demand and Supply .......................2-18 2.2.2 Equilibrium .................................................2-20 2.2.3 Efficiency of Equilibrium ............................2-22 2.3 Changes in Supply and Demand..........2-22 2.3.1 Changes in Demand....................................2-22 2.3.2 Changes in Supply.......................................2-23 2.4 Elasticities................................................2-27 2.4.1 Elasticity of Dem and...................................2-27 2.4.2 Elasticity of Supply .....................................2-30 2.5 Comparative Statics ...............................2-30 2.5.1 Supply and Demand Changes .....................2-30 2.6 T rade .........................................................2-32 2.6.1 Production Possibilities Frontier ................2-32 2.6.2 Comparative and Absolute Advantage ........2-36 2.6.3 Factors and Production...................2-38 2.6.4 International Trade......................2-39 3 THE US ECONOMY.....................................3-41 3.1.1 Basic Demographics ..................3-41 3.1.2 Education... ...............................3-47 3.1.3 Households and Consumption ..........3-49 3.1.4 Production................................................3-56 3.1.5 G overnm ent................................................3-65 3.1.6 T rade...........................................................3-73 3.1.7 Fluctuations ................................................3-76 4 PRODUCER THEORY.................................4-79 4.1 The Competitive Firm............................4-79 4.1.1 Types of Firms........................................4-79 4.1.2 Production Functions....................4-81 4.1.3 Profit M axim ization ....................................4-85 4.1.4 The Shadow Value.......................................4-91 4.1.5 Input Dem and.............................................4-92 4.1.6 M yriad Costs ...............................................4-95 4.1.7 Dynamic Firm Behavior ..............................4-97 4.1.8 Economies of Scale and Scope ..................4-100 4.2 Perfect Competition Dynamics..........4-104 4.2.1 Long-run Equilibrium...............................4-104 4.2.2 Dynamics with Constant Costs...... ...4-105 4.2.3 General Long-run Dynamics...............4-109 4.3 Inestment..............................4-114 4.1 Present value............4-114 4.3.2 Investment.................................4-118 4.3.3 Investment Under Uncertainty............4-120 4.3.4 Resource Extraction.......................4-125 4.3.5 A Time to Harvest...........................4-127 4.3.6 Collectibles.................................4-130 4.3.7 Summer Wheat...............................4-135 5 CONSUMER THEORY ............................. 5-139 5.1 5.1.1 5.1.2 5.1.3 5.1.4 5.1.5 5.2 5.2.1 5.2.2 5.2.3 5.2.4 5.2.5 5.2.6 5.2.7 5.2.8 5.2.9 6 6.1 6.1.1 6.1.2 6.1.3 6.2 6.2.1 6.2.2 6.2.3 6.2.4 6.2.5 6.3 6.3.1 6.3.2 6.3.3 6.3.4 6.3.5 6.3.6 6.4 6.4.1 6.4.2 6.4.3 6.4.4 6.5 6.5.1 6.5.2 6.5.3 6.5-4 6.5.5 6.5.6 6.5.7 6.5.8 6.6 6.6.1i 6.6.2 6.6.3 7 7.1 7.1.1 7.1.2 7-1-3 7-1-4 7-1-5 Utility Maximization ........................... 5-139 Budget or Feasible Set...............5-140 Isoquants ................................................. 5-143 Examples.................................................. 5-148 Substitution Effects.................5-151 Income Effects.......................................... 5-155 Additional Considerations................. 5-158 Corner Solutions...................5-158 Labor Supply ............................................ 5-160 Compensating Differentials...........5-164 Urban Real Estate Prices.............5-165 Dynamic Choice ....................................... 5-169 Risk .......................................................... 5-174 Search ...................................................... 5-178 Edgeworth Box....................5-181 General Equilibrium................5-188 MARKET IMPERFECTIONS................... 6-195 Taxes....................................................... 6-195 Effects of Taxes....................6-195 Incidence of Taxes ..................6-199 Excess Burden of Taxation ....................... 6-200 Price Floors and Ceilings ................... 6-202 Basic Theory............................................. 6-203 Long- and Short-run Effects ...........6-207 Political Motivations ................ 6-209 Price Supports .......................................... 6-210 Quantity Restrictions and Quotas......6-211 Externalities ......................................... 6-213 Private and Social Value, Cost.........6-214 Pigouvian Taxes.......................................6-217 Q uotas ......................................................6-2 18 Tradable Permits and Auctions........6-219 Coasian Bargaining.................6-220 Fishing and Extinction ............................. 6-221 Public Goods......................................... 6-226 Examples.................................................. 6-226 Free-Riders .............................................. 6-227 Provision with Taxation ........................... 6-229 Local Public Goods.................6-230 Monopoly............................................... 6-232 Sources of Monopoly................6-232 Basic Analysis........................................... 6-233 Effect of Taxes .......................................... 6-236 Price Discrimination................6-237 Welfare Effects ......................................... 6-240 Two-Part Pricing...................6-240 Natural Monopoly..................6-241 Peak Load Pricing..................6-242 Information .......................................6-245 Market for Lemons...........................6-245 Myerson-Satterthwaite Theorem.........6-246 Signaling...................................... 6-248 STRATEGIC BEHAVIOR .................. 7-251 Games..................................... 7-251 Matrix Games.............................. 7-251 Nash Equilibrium.......................... 7-255 Mixed Strategies..............................7-257 Examples................................... 7-262 Two Period Games.......................... 7-265 McAfee: Introduction to Economic Analysis, http://www.introecon.com, July 24, 2006 jv 7.1.6 Subgame Perfection.................7-266 7.6 Auctions..................................................7-295 7.1.7 Supergam es .............................................. 7-268 7.6.1 English Auction.........................................7-295 7.1.8 The Folk Theorem..................7-269 7.6.2 Sealed-bid Auction..................7-296 7.2 Cournot Oligopoly................7-270 7.6.3 Dutch Auction...........................................7-298 7.2.1 Equilibrium ..............................................7-271 7.6.4 Vickrey Auction ....................7-299 7.2.2 Industry Performance...............7-272 7.6.5 Winner's Curse..........................................7-301 7.3 Search and Price Dispersion........7-274 7.6.6 Linkage .....................................................7-303 7.3.1 Simplest Theory. ....................7-275 7.6.7 Auction Design..........................................7-304 7.3.2 Industry Performance...............7-277 7.7 Antitrust .................................................7-306 7.7.1 Sherm an A ct .............................................7-306 7.4 Hotelling Model..................7-279 7.7.1 hran Act... ...................7-306 7.4.1 Types of Differentiation............................ 7-279 7.7.2 PceFiongct .......................7-308 7.4.2 The Standard Model ................7-280e-Fixing..............................................7-309 7.4.3 The Circle Model..................1......7-280 7.5 Agency Theory7-283 8 INDEX.....................................................8-315 7.5.1 Sim ple M odel............................................ 7-284 8.1 List of Figures .......................................8-315 7.5.2 Cost of Providing Incentives ..................... 7-286 8.2 Index ...........................8-317 7.5.3 Selection of Agent..................7-287 7.5.4 M ulti-tasking ............................................ 7-288 7.5.5 Multi-tasking without Homogeneity.........7-292 McAfee: Introduction to Economic Analysis, http://www.introecon.com, July 24, 2006 V 1 What is Economics? Economics studies the allocation of scarce resources among people - examining what goods and services wind up in the hands of which people. Why scarce resources? Absent scarcity, there is no significant allocation issue. All practical, and many impractical, means of allocating scarce resources are studied by economists. Markets are an important means of allocating resources, so economists study markets. Markets include stock markets like the New York Stock Exchange, commodities markets like the Chicago Mercantile, but also farmer's markets, auction markets like Christie's or Sotheby's (made famous in movies by people scratching their noses and inadvertently purchasing a Ming vase) or eBay, or more ephemeral markets, such as the market for music CDs in your neighborhood. In addition, goods and services (which are scarce resources) are allocated by governments, using taxation as a means of acquiring the items. Governments may be controlled by a political process, and the study of allocation by the politics, which is known as political economy, is a significant branch of economics. Goods are allocated by certain means, like theft, deemed illegal by the government, and such allocation methods nevertheless fall within the domain of economic analysis; the market for marijuana remains vibrant despite interdiction by the governments of most nations. Other allocation methods include gifts and charity, lotteries and gambling, and cooperative societies and clubs, all of which are studied by economists. Some markets involve a physical marketplace. Traders on the New York Stock Exchange get together in a trading pit. Traders on eBay come together in an electronic marketplace. Other markets, which are more familiar to most of us, involve physical stores that may or may not be next door to each other, and customers who search among the stores, purchasing when the customer finds an appropriate item at an acceptable price. When we buy bananas, we don't typically go to a banana market and purchase from one of a dozen or more banana sellers, but instead go to a grocery store. Nevertheless, in buying bananas, the grocery stores compete in a market for our banana patronage, attempting to attract customers to their stores and inducing them to purchase bananas. Price - exchange of goods and services for money - is an important allocation means, but price is hardly the only factor even in market exchanges. Other terms, such as convenience, credit terms, reliability, and trustworthiness are also valuable to the participants in a transaction. In some markets such as 36 inch Sony WEGA televisions, one ounce bags of Cheetos, or Ford Autolite spark plugs, the products offered by distinct sellers are identical, and for such products, price is usually the primary factor considered by buyers, although delivery and other aspects of the transaction may still matter. For other products, like restaurant meals, camcorders by different manufacturers, or air travel on distinct airlines, the products differ to some degree, and thus the qualities of the product are factors in the decision to purchase. Nevertheless, different products may be considered to be in a single market if the products are reasonable substitutes, and we can consider a "quality-adjusted" price for these different goods. McAfee: Introduction to Economic Analysis, http://www.introecon.com, July 24, 2006 1.1 Economic analysis is used in many situations. When British Petroleum sets the price for its Alaskan crude oil, it uses an estimated demand model, both for gasoline consumers and also for the refineries to which BP sells. The demand for oil by refineries is governed by a complex economic model used by the refineries and BP estimates the demand by refineries by estimating the economic model used by refineries. Economic analysis was used by experts in the antitrust suit brought by the U.S. Department of Justice both to understand Microsoft's incentive to foreclose (eliminate from the market) rival Netscape and consumer behavior in the face of alleged foreclosure. Stock market analysts use economic models to forecast the profits of companies in order to predict the price of their stocks. When the government forecasts the budget deficit or considers a change in environmental regulations, it uses a variety of economic models. This book presents the building blocks of the models in common use by an army of economists thousands of times per day. 1.1.1 Normative and Positive Theories Economic analysis is used for two main purposes. The first is a scientific understanding of how allocations of goods and services - scarce resources - are actually determined. This is a positive analysis, analogous to the study of electromagnetism or molecular biology, and involves only the attempt to understand the world around us. The development of this positive theory, however, suggests other uses for economics. Economic analysis suggests how distinct changes in laws, rules and other government interventions in markets will affect people, and in some cases, one can draw a conclusion that a rule change is, on balance, socially beneficial. Such analyses combine positive analysis - predicting the effects of changes in rules - with value judgments, and are known as normative analyses. For example, a gasoline tax used to build highways harms gasoline buyers (who pay higher prices), but helps drivers (who face fewer potholes and less congestion). Since drivers and gasoline buyers are generally the same people, a normative analysis may suggest that everyone will benefit. This type of outcome, where everyone is made better off by a change, is relatively uncontroversial. In contrast, cost-benefit analysis weighs the gains and losses to different individuals and suggests carrying out changes that provide greater benefits than harm. For example, a property tax used to build a local park creates a benefit to those who use the park, but harms those who own property (although, by increasing property values, even non-users obtain some benefits). Since some of the taxpayers won't use the park, it won't be the case that everyone benefits on balance. Cost-benefit analysis weighs the costs against the benefits. In the case of the park, the costs are readily monetized (turned into dollars), because the costs to the tax-payers are just the amount of the tax. In contrast, the benefits are much more challenging to estimate. Conceptually, the benefits are the amount the park users would be willing to pay to use the park if the park charged admission. However, if the park doesn't charge admission, we would have to estimate willingness-to-pay. In principle, the park provides greater benefits than costs if the benefits to the users exceed the losses to the taxpayers. However, the park also involves transfers from one group to another. Welfare analysis provides another approach to evaluating government intervention into markets. Welfare analysis posits social preferences and goals, like helping the poor. Generally a welfare analysis involves performing a cost-benefit analysis taking account McAfee: Introduction to Economic Analysis, http://www.introecon.com, July 24, 2006 1.2 not just of the overall gains and losses, but also weighting those gains and losses by their effects on other social goals. For example, a property tax used to subsidize the opera might provide more value than costs, but the bulk of property taxes are paid by lower and middle income people, while the majority of opera-goers are rich. Thus, the opera subsidy represents a transfer from relatively low income people to richer people, which is not consistent with societal goals of equalization. In contrast, elimination of sales taxes on basic food items like milk and bread generally has a relatively greater benefit to the poor, who spend a much larger percentage of their income on food, than to the rich. Thus, such schemes may be considered desirable not so much for their overall effects but for their redistribution effects. Economics is helpful not just in providing methods for determining the overall effects of taxes and programs, but also the incidence of these taxes and programs, that is, who pays, and who benefits. What economics can't do, however, is say who ought to benefit. That is a matter for society at large to decide. 1.1.2 Opportunity Cost Economists use the idea of cost in a slightly quirky way that makes sense once you think about it, and we use the term opportunity cost to remind you occasionally of our idiosyncratic notion of cost. For an economist, the cost of something is not just the cash payment, but all of the value given up in the process of acquiring the thing. For example, the cost of a university education involves tuition, and text book purchases, and also the wages that would have been earned during the time at university, but were not. Indeed, the value of the time spent in acquiring the education - how much enjoyment was lost - is part of the cost of education. However, some "costs" are not opportunity costs. Room and board would not generally be a cost because, after all, you are going to be living and eating whether you are in university or not. Room and board are part of the cost of an education only insofar as they are more expensive than they would be otherwise. Similarly, the expenditures on things you would have otherwise done - hang-gliding lessons, a trip to Europe - represent savings. However, the value of these activities has been lost while you are busy reading this book. The concept of opportunity cost can be summarized by a definition: The opportunity cost is the value of the best foregone alternative. This definition captures the idea that the cost of something is not just its monetary cost but also the value of what you didn't get. The opportunity cost of spending $17 on a CD is what you would have done with the $17 instead, and perhaps the value of the time spent shopping. The opportunity cost of a puppy includes not just the purchase price of the puppy, but also the food, veterinary bills, carpet cleaning, and the value of the time spent dealing with the puppy. A puppy is a good example, because often the purchase price is a negligible portion of the total cost of ownership. Yet people acquire puppies all the time, in spite of their high cost of ownership. Why? The economic view of the world is that people acquire puppies because the value they expect to get exceeds the opportunity cost. That is, they acquire a puppy when the value of a puppy is higher than the value of what is foregone by the acquisition of a puppy. Even though opportunity costs include lots of non-monetary costs, we will often monetize opportunity costs, translating the costs into dollar terms for comparison McAfee: Introduction to Economic Analysis, http://www.introecon.com, July 24, 2006 1.3 purposes. Monetizing opportunity costs is clearly valuable, because it gives a means of comparison. What is the opportunity cost of 30 days in jail? It used to be that judges occasionally sentenced convicted defendants to "thirty days or thirty dollars," letting the defendant choose the sentence. Conceptually, we can use the same idea to find out the value of 30 days in jail. Suppose you would choose to pay a fine of $750 to avoid the thirty days in jail, but wouldn't pay $1,ooo and instead would choose time in the slammer. Then the value of the thirty day sentence is somewhere between $750 and $1ooo. In principle, there exists a price where at that price you pay the fine, and at a penny more you go to jail. That price - at which you are just indifferent to the choice - is the monetized or dollar cost of the jail sentence. The same idea as choosing the jail sentence or the fine justifies monetizing opportunity costs in other contexts. For example, a gamble has a certainty equivalent, which is the amount of money that makes one indifferent to choosing the gamble versus the certain amount. Indeed, companies buy and sell risk, and much of the field of risk management involves buying or selling risky items to reduce overall risk. In the process, risk is valued, and riskier stocks and assets must sell for a lower price (or, equivalently, earn a higher average return). This differential is known as a risk premium, and it represents a monetization of the risk portion of a risky gamble. Home buyers considering various available houses are presented with a variety of options, such as one or two story, building materials like brick or wood, roofing materials, flooring materials like wood or carpet, presence or absence of swimming pools, views, proximity to parks, and so on. The approach taken to valuing these items is known as hedonic pricing, and corresponds to valuing each item separately - what does a pool add to value on average? - and then summing the value of the components. The same approach is used to value old cars, making adjustments to a base value for the presence of options like leather interior, CD changer, and so on. Again, such a valuation approach converts a bundle of disparate attributes into a monetary value. The conversion of costs into dollars is occasionally controversial, and nowhere is it more controversial than in valuing human life. How much is your life worth? Can it be converted into dollars? A certain amount of insight into this question can be gleaned by thinking about risks. Wearing seatbelts and buying optional safety equipment reduce the risk of death by a small but measurable amount. Suppose a $400 airbag option reduces the overall risk of death by o.oi%. If you are indifferent to buying the option, you have implicitly valued the probability of death at $400 per o.oi%, or $40,000 per 1%, or around $4,000,000 per life. Of course, you may feel quite differently about a 0.01% chance of death than a risk ten thousand times greater, which would be a certainty. But such an approach provides one means of estimating the value of the risk of death - an examination what people will, and will not, pay to reduce that risk. Opportunity cost - the value of the best foregone alternative - is a basic building block of economic analysis. The conversion of costs into dollar terms, while sometimes controversial, provides a convenient means of comparing costs. McAfee: Introduction to Economic Analysis, http://www.introecon.com, July 24, 2006 1.4 1.1.3 Economic Reasoning and Analysis What this country needs is some one-armed economists. -Harry S Truman Economic reasoning is rather easy to satirize. One might want to know, for instance, what the effect of a policy change - a government program to educate unemployed workers, an increase in military spending, or an enhanced environmental regulation - will be on people and their ability to purchase the goods and services they desire. Unfortunately, a single change may have multiple effects. As an absurd and tortured example, government production of helium for (allegedly) military purposes reduces the cost of children's birthday balloons, causing substitution away from party hats and hired clowns. The reduction in demand for clowns reduces clowns' wages and thus reduces the costs of running a circus. This cost reduction increases the number of circuses, thereby forcing zoos to lower admission fees to compete with circuses. Thus, were the government to stop subsidizing the manufacture of helium, the admission fee of zoos would likely rise, even though zoos use no helium. This example is superficially reasonable, although the effects are miniscule. To make any sense at all of the effects of a change in economic conditions, it is helpful to divide up the effect into pieces. Thus, we will often look at the effects of a change "other things equal," that is, assuming nothing else changed. This isolates the effect of the change. In some cases, however, a single change can lead to multiple effects; even so, we will still focus on each effect individually. A gobbledygook way of saying "other things equal" is to use Latin and say "ceteris paribus." Part of your job as a student is to learn economic jargon, and that is an example. Fortunately, there isn't too much jargon. We will make a number of assumptions that you may not find very easy to believe. Not all of the assumptions are required for the analysis, and instead merely simplify the analysis. Some, however, are required but deserve an explanation. There is a frequent assumption that the people we will talk about seem exceedingly selfish relative to most people we know. We model the choices that people make, assuming that they make the choice that is best for them. Such people - the people in the models as opposed to real people - are known occasionally as "homo economicus." Real people are indubitably more altruistic than homo economicus, because they couldn't be less: homo economicus is entirely selfish. (The technical term is acting in one's self-interest.) That doesn't necessarily invalidate the conclusions drawn from the theory, however, for at least four reasons: " People often make decisions as families or households rather than individuals, and it may be sensible to consider the household as the "consumer." That households are fairly selfish is more plausible perhaps than individuals being selfish. * Economics is pretty much silent on why consumers want things. You may want to make a lot of money so that you can build a hospital or endow a library, which would be altruistic things to do. Such motives are broadly consistent with self- interested behavior. * Corporations are often required to serve their shareholders by maximizing the share value, inducing self-interested behavior on the part of the corporation. Even if corporations had no legal responsibility to act in the financial interest of McAfee: Introduction to Economic Analysis, http://www.introecon.com, July 24, 2006 1-5 their shareholders, capital markets may force them to act in the self-interest of the shareholders in order to raise capital. That is, people choosing investments that generate a high return will tend to force corporations to seek a high return. " There are many good, and some not-so-good, consequences of people acting in their own self-interest, which may be another reason to focus on self-interested behavior. Thus, while there are limits to the applicability of the theory of self-interested behavior, it is a reasonable methodology for attempting a science of human behavior. Self-interested behavior will often be described as "maximizing behavior," where consumers maximize the value they obtain from their purchases, and firms maximize their profits. One objection to the economic methodology is that people rarely carry out the calculations necessary to literally maximize anything. However, that is not a sensible objection to the methodology. People don't carry out the physics calculations to throw a baseball or thread a needle, either, and yet they accomplish these tasks. Economists often consider that people act "as if" they maximize an objective, even though no calculations are carried out. Some corporations in fact use elaborate computer programs to minimize costs or maximize their profits, and the entire field of operations research is used to create and implement such maximization programs. Thus, while individuals don't carry out the calculations, some companies do. A good example of economic reasoning is the sunk cost fallacy. Once one has made a significant non-recoverable investment, there is a psychological tendency to invest more even when the return on the subsequent investment isn't worthwhile. France and Britain continued to invest in the Concorde (a supersonic aircraft no longer in production) long after it became clear that the project would generate little return. If you watch a movie to the end, long after you become convinced that it stinks, you have exhibited the sunk cost fallacy. The fallacy is the result of an attempt to make an investment that has gone bad turn out to be good, even when it probably won't. The popular phrase associated with the sunk cost fallacy is "throwing good money after bad." The fallacy of sunk costs arises because of a psychological tendency to try to make an investment pay off when something happens to render it obsolete. It is a mistake in many circumstances. The fallacy of sunk costs is often thought to be an advantage of casinos. People who lose a bit of money gambling hope to recover their losses by gambling more, with the sunk "investment" in gambling inducing an attempt to make the investment pay off. The nature of most casino gambling is that the house wins on average, which means the average gambler (and even the most skilled slot machine or craps player) loses on average. Thus, for most, trying to win back losses is to lose more on average. The way economics is performed is by a proliferation of mathematical models, and this proliferation is reflected in this book. Economists reason with models. Models help by removing extraneous details from a problem or issue, letting one analyze what remains more readily. In some cases the models are relatively simple, like supply and demand. In other cases, the models are relatively complex (e.g. the over-fishing model of Section 6.3.6). In all cases, the models are the simplest model that lets us understand the question or phenomenon at hand. The purpose of the model is to illuminate McAfee: Introduction to Economic Analysis, http://www.introecon.com, July 24, 2006 1.6 connections between ideas. A typical implication of a model is "when A increases, B falls." This "comparative static" prediction lets us see howA affects B, and why, at least in the context of the model. The real world is always much more complex than the models we use to understand the world. That doesn't make the model useless, indeed, exactly the opposite. By stripping out extraneous detail, the model represents a lens to isolate and understand aspects of the real world. Finally, one last introductory warning before we get started. A parody of economists talking is to add the word marginal before every word. Marginal is just economist's jargon for "the derivative of." For example, marginal cost is the derivative of cost; marginal value is the derivative of value. Because introductory economics is usually taught to students who have not yet studied calculus or can't be trusted to remember even the most basic elements of it, economists tend to avoid using derivatives and instead talk about the value of the next unit purchased, or the cost of the next unit, and describe that as the marginal value or cost. This book uses the term marginal frequently because one of the purposes of the book is to introduce the necessary jargon so that you can read more advanced texts or take more advanced classes. For an economics student not to know the word marginal would be akin to a physics student not knowing the word mass. The book minimizes jargon where possible, but part of the job of a principles student is to learn the jargon, and there is no getting around that. McAfee: Introduction to Economic Analysis, http://www.introecon.com, July 24, 2006 1.7 2 Supply and Demand Supply and demand are the most fundamental tools of economic analysis. Most applications of economic reasoning involve supply and demand in one form or another. When prices for home heating oil rise in the winter, usually the reason is that the weather is colder than normal and as a result, demand is higher than usual. Similarly, a break in an oil pipeline creates a short-lived gasoline shortage, as occurred in the Midwest in the year 2000, which is a reduction in supply. The price of DRAM, or dynamic random access memory, used in personal computers falls when new manufacturing facilities begin production, increasing the supply of memory. This chapter sets out the basics of supply and demand, introduces equilibrium analysis, and considers some of the factors that influence supply and demand and the effects of those factors. In addition, quantification is introduced in the form of elasticities. Dynamics are not considered, however, until Chapter 4, which focuses on production, and Chapter 5 introduces a more fundamental analysis of demand, including a variety of topics such as risk. In essence, this is the economics "quickstart" guide, and we will look more deeply in the subsequent chapters. 2.1 Supply and Demand 2.1.1 Demand and Consumer Surplus Eating a French fry makes most people a little bit happier, and we are willing to give up something of value - a small amount of money, a little bit of time - to eat one. What we are willing to give up measures the value - our personal value - of the French fry. That value, expressed in dollars, is the willingness to pay for French fries. That is, if you are willing to give up three cents for a single French fry, your willingness to pay is three cents. If you pay a penny for the French fry, you've obtained a net of two cents in value. Those two cents - the difference between your willingness to pay and the amount you do pay - is known as consumer surplus. Consumer surplus is the value to a consumer of consumption of a good, minus the price paid. The value of items - French fries, eyeglasses, violins - is not necessarily close to what one has to pay for them. For people with bad vision, eyeglasses might be worth ten thousand dollars or more, in the sense that if eyeglasses and contacts cost $io,ooo at all stores, that is what one would be willing to pay for vision correction. That one doesn't have to pay nearly that amount means that the consumer surplus associated with eyeglasses is enormous. Similarly, an order of French fries might be worth $3 to a consumer, but because French fries are available for around $1, the consumer obtains a surplus of $2 in the purchase. How much is a second order of French fries worth? For most of us, that first order is worth more than the second one. If a second order is worth $2, we would still gain from buying it. Eating a third order of fries is worth less still, and at some point we're unable or unwilling to eat any more fries even when they are free, which implies that at some point the value of additional French fries is zero. We will measure consumption generally as units per period of time, e.g. French fries consumed per month. McAfee: Introduction to Economic Analysis, http://www.introecon.com, July 24, 2006 2.8 Many, but not all, goods have this feature of diminishing marginal value - the value of the last unit consumed declines as the number consumed rises. If we consume a quantity q, it implies the marginal value v(q) falls as the number of units rise.' An example is illustrated in Figure 2-1. Here the value is a straight line, declining in the number of units. v(qo), p p Figure 2-1: The Demand Curve Demand need not be a straight line, and indeed could be any downward-sloping curve. Contrary to the usual convention, demand gives the quantity chosen for any given price off the horizontal axis, that is, given the value p on the vertical axis, the corresponding value qo on the horizontal axis is the quantity the consumer will purchase. It is often important to distinguish the demand curve itself - the entire relationship between price and quantity demanded - from the quantity demanded. Typically, "demand" refers to the entire curve, while "quantity demanded" is a point on the curve. Given a price p, a consumer will buy those units with v(q) >p, since those units are worth more than they cost. Similarly, a consumer should not buy units for which v(q)
1), the optimal monopoly
quantity is essentially zero, and in any event would be no more than one molecule of the
product.
76 Abba Lerner, 1903-1982. Note that 1 Gm dqc susdi h
p(qm) /PCqm) /P
derivation.
McAfee: Introduction to Economic Analysis, http://www.introecon.com, July 24, 2006
6-235
In addition, the effects of monopoly are related to the elasticity of demand. If demand is
very elastic, the effect of monopoly on prices is quite limited. In contrast, if the demand
is relatively inelastic, monopolies will increase prices by a large margin.
We can rewrite the formula to obtain
8-1 c(q,).
Thus, a monopolist marks up marginal cost by the factor ,at least when E>1. This
8-1
formula is sometimes used to justify a "fixed markup policy," which means a company
adds a constant percentage markup to its products. This is an ill-advised policy not
justified by the formula, because the formula suggests a markup which depends on the
demand for the product in question and thus not a fixed markup for all products a
company produces.
6.5.3 Effect of Taxes
A tax imposed on a seller with monopoly power performs differently than a tax imposed
on a competitive industry. Ultimately a perfectly competitive industry must pass on all
of a tax to consumers, because in the long-run the competitive industry earns zero
profits. In contrast, a monopolist might absorb some portion of a tax even in the long-
run.
To model the effect of taxes on a monopoly, consider a monopolist who faces a tax rate t
per unit of sales. This monopolist earns
t= p(q)q -c(q)-tq.
The first order condition for profit maximization yields
0_ p(qm )+q m m)-c'(qm )-t .
89
Viewing the monopoly quantity as a function of t, we obtain:
dqm _ 1
dt 2p'(q)+qwp"(qm)-c"(q)m)
with the sign following from the second order condition for profit maximization. In
addition, the change in price satisfies
py(qdqm _ p'(qm) >0.
dt 2p'(qm) +qmp"(qm) -c"(qm)
McAfee: Introduction to Economic Analysis, http://www.introecon.com, July 24, 2006
6.236
Thus, a tax causes a monopoly to increase its price. In addition, the monopoly price
rises by less than the tax if p'(qm) dqm <1, or
mdt
p'(qm )+qmp"(qm )- c"(qm)< 0.
This condition need not be true, but is a standard regularity condition imposed by
assumption. It is true for linear demand and increasing marginal cost. It is false for
constant elasticity of demand, s>i (which is the relevant case, for otherwise the second
order conditions fail) and constant marginal cost. In the latter case (constant elasticity
and marginal cost), a tax on a monopoly increases price by more than the amount of the
tax.
6.5.3.1 (Exercise) Use a revealed preference argument to show that a per unit tax
imposed on a monopoly causes the quantity to fall. That is, hypothesize
quantities qa before the tax, and qa after the tax, and show that two facts - the
before tax monopoly preferred qi to qa and the taxed monopoly made higher
profits from qa together imply the q< qa.
6.5.3.2 (Exercise) When both demand and supply have constant elasticity, use the
results of 6.5.2.3 (Exercise) to compute the effect of a proportional tax (i.e. a
portion of the price paid to the government).
6.5.4 Price Discrimination
Pharmaceutical drugs for sale in Mexico are generally priced substantially below their
U.S. counterparts. Pharmaceutical drugs in Europe are also cheaper than in the U.S.,
although not as inexpensive as in Mexico, with Canadian prices usually between the U.S.
and European prices. (The comparison is between identical drugs produced by the same
manufacturer.)
Pharmaceutical drugs differ in price across countries primarily because demand
conditions vary. The formula
8
pfm )= E1 c'(qm ).
shows that a monopoly seller would like to charge a higher markup over marginal cost to
customers with less elastic demand than to customers with more elastic demand,
because is a decreasing function of 3, for s>1i. Charging different prices for the
8 -1
same product to different customers is known as price discrimination. In business
settings, it is sometimes known as value-based pricing, which is a more palatable term
to tell to customers.
Computer software vendors often sell a "student" version of their software, usually at
substantially reduced prices, and requiring proof of being a student to qualify for the
lower price. Such student discounts are examples of price discrimination, and students
McAfee: Introduction to Economic Analysis, http://www.introecon.com, July 24, 2006
6.237
have more elastic demand than business users. Similarly, the student and senior citizen
discounts at movies and other venues sell the same thing - a ticket to the show - for
different prices, and thus qualify as price discrimination.
In order for a seller to price-discriminate, the seller must be able to
" identify (approximately) the demand of groups of customers
" prevent arbitrage
Arbitrage is also known as "buying low and selling high," and represents the act of being
an intermediary. Since price discrimination requires charging one group a higher price
than another, there is potentially an opportunity for arbitrage, arising from members of
the low price group buying at the low price and selling at the high price. If the seller
can't prevent arbitrage, arbitrage essentially converts a two-price system to sales at the
low price.
Why offer student discounts at the movies? You already know the answer to this -
students have lower incomes on average than others, and lower incomes translate into a
lower willingness to pay for normal goods. Consequently a discount to a student makes
sense from a demand perspective. Arbitrage can be mostly prevented by requiring a
student identification card to be presented. Senior citizen discounts are a bit more
subtle. Generally seniors aren't poorer than other groups of customers (in the United
States, at least). However, seniors have more free time, and thus are able to substitute
to matinee showings77 or drive to more distant locations should those offer discounts.
Thus seniors have relatively elastic demand more because of their ability to substitute
than because of their income.
Airlines commonly price discriminate, using "Saturday night stay-overs" and other
devices. To see that such charges represent price discrimination, consider a passenger
who lives in Dallas but needs to spend Monday through Thursday in Los Angeles two
weeks in a row. This passenger could buy two round-trip tickets:
Trip One:
First Monday: Dallas -> Los Angeles
First Friday: Los Angeles -> Dallas
Trip Two:
Second Monday: Dallas -> Los Angeles
Second Friday: Los Angeles -> Dallas
At the time of this writing, the approximate combined cost of these two flights was
US$2,000. In contrast, another way of arranging exactly the same travel is to have two
round-trips, one of which originates in Dallas, while the other originates in Los Angeles:
77 Matinee showings are those early in the day, which are usually discounted. These discounts are not
price discrimination because a show at noon isn't the same product as a show in the evening.
McAfee: Introduction to Economic Analysis, http://www.introecon.com, July 24, 2006
6.238
Trip One:
First Monday: Dallas -> Los Angeles
Second Friday: Los Angeles -> Dallas
Trip Two:
First Friday: Los Angeles -> Dallas
Second Monday: Dallas -> Los Angeles
This pair of round trips involves exactly the same travel as the first pair, but costs less
than $500 for both (at the time of this writing). The difference is that the second pair
involves staying over Saturday night for both legs, and that leads to a major discount for
most U.S. airlines. (American Airlines quoted the fares.)
How can airlines price discriminate? There are two major groups of customers:
business travelers and leisure travelers. Business travelers have the higher willingness
to pay overall, and the nature of their trips tends to be that they come home for the
weekend. In contrast, a leisure traveler will usually want to be away for a weekend, so a
weekend stay-over is an indicator of a leisure traveler. It doesn't work perfectly as an
indicator - some business travelers must be away for the weekend - but it is sufficiently
correlated with leisure travel that it is profitable for the airline to price discriminate.
These examples illustrate an important distinction. Senior citizen and student discounts
are based on the identity of the buyer, and qualifying for the discount requires showing
an identity card. In contrast, airline price discrimination is not based on the identity of
the buyer but on the choices by the buyer. The former is known as direct price
discrimination, while the latter is known as indirect price discrimination.78
Two common examples of indirect price discrimination are coupons and quantity
discounts. Coupons offer discounts for products and are especially common in grocery
stores, where they are usually provided in a newspaper section available free at the front
of the store. Coupons discriminate on the basis of the cost of time. It takes time to find
the coupons for the products one is interested in buying, and thus those with a high
value of time won't find it worthwhile spending twenty minutes to save $5 (effectively a
$15 per hour return), while those with a low value of time will find that return
worthwhile. Since those with a low value of time tend to be more price sensitive (more
elastic demand), coupons offer a discount available to all but used primarily by
customers with a more elastic demand, and thus increase the profits of the seller.
Quantity discounts are discounts for buying more. Thus, the large size of milk, laundry
detergent and other items often cost less per unit than smaller sizes, and the difference
is greater than the savings on packaging costs. In some cases, the larger sizes entail
greater packaging costs; some manufacturers "band together" individual units, incurring
additional costs to create a larger size which is then discounted. Thus, the "twenty-four
pack" of paper towels sells for less per roll than the individual rolls; such large volumes
appeal primarily to large families, who are more price-sensitive on average.
78 The older and incoherent language for these concepts called direct price discrimination "third degree
price discrimination," while indirect price discrimination was called second degree price discrimination.
In the older language, first degree price discrimination meant perfect third degree price discrimination.
McAfee: Introduction to Economic Analysis, http://www.introecon.com, July 24, 2006
6.239
6.5.5 Welfare Effects
Is price discrimination a good thing, or a bad thing? It turns out that there is no
definitive answer to this question. Instead, it depends on circumstances. We illustrate
this conclusion with a pair of exercises.
6.5.5.1 (Exercise) Let marginal cost be zero for all quantities. Suppose there are two
equal-sized groups of customers, group 1 with demand q(p)=12-p, group 2 with
demand q(p)=8-p. Show that a non-discriminating monopolist charges a price
of 5 and the discriminating monopolist charges group 1 the price 6 and group 2
the price 4. Then calculate the gains from trade, with discrimination and
without, and show that price discrimination reduces the gains from trade.
This exercise illustrates a much more general proposition: if a price-discriminating
monopolist produces less than a non-discriminating monopolist, then price
discrimination reduced welfare. This proposition has an elementary proof. Consider
the price discriminating monopolist's sales, and then allow arbitrage. The arbitrage
increases the gains from trade, since every transaction has gains from trade. Arbitrage,
however, leads to a common price like that charged by a non-discriminating monopolist.
Thus, the only way price discrimination can increase welfare is if it leads a seller to sell
more output than she would otherwise. This is possible, as the next exercise shows.
6.5.5.2 (Exercise) Let marginal cost be zero for all quantities. Suppose there are two
equal-sized groups of customers, group 1 with demand q(p)=12-p, group 2 with
demand q(p)=4-p. Show that a non-discriminating monopolist charges a price
of 6 and the discriminating monopolist charges group 1 the price 6 and group 2
the price 2. Then calculate the gains from trade, with discrimination and
without, and show that price discrimination increases the gains from trade.
In this exercise, we see that price discrimination that brings in a new group of customers
may increase the gains from trade. Indeed, this example involves a Pareto
improvement: the seller and group 2 are better off, and group 1 no worse off, than
without price discrimination. (A Pareto improvement requires that no one is worse off
and at least one person is better off.)
Whether price discrimination increases the gains from trade overall depends on
circumstances. However, it is worth remembering that people with lower incomes tend
to have more elastic demand, and thus get lower prices under price discrimination than
absent price discrimination. Consequently, a ban on price discrimination tends to hurt
the poor and benefit the rich no matter what the overall effect.
6-5-6 Two-Part Pricing
A common form of price discrimination is known as two-part pricing. Two-part pricing
usually involves a fixed charge and a marginal charge, and thus offers an ability for a
seller to capture a portion of the consumer surplus. For example, electricity often comes
with a fixed price per month and then a price per kilowatt-hour, which is two-part
pricing. Similarly, long distance and cellular telephone companies charge a fixed fee per
month, with a fixed number of "included" minutes, and a price per minute for additional
McAfee: Introduction to Economic Analysis, http://www.introecon.com, July 24, 2006
6.240
minutes. Such contracts really involve three parts rather than two-parts, but are similar
in spirit.
From the seller's perspective, the ideal two-part price is to charge marginal cost plus a
fixed charge equal to the customer's consumer surplus, or perhaps a penny less. By
setting price equal to marginal cost, the seller maximizes the gains from trade. By
setting the fixed fee equal to consumer surplus, the seller captures the entire gains from
trade. This is illustrated in Figure 6-21.
qc
Figure 6-21: Two-Part Pricing
6.5.7 Natural Monopoly
A natural monopoly arises when a single firm can efficiently serve the entire market
because average costs are lower with one firm than with two firms. An example is
illustrated in Figure 6-22. In this case, the average total cost of a single firm is lower
than if two firms operate, splitting the output between them. The monopolist would like
to price at pm, which maximizes profits.79
79 The monopoly price may or may not be sustainable. A monopoly price is not sustainable if it would
lead to entry, thereby undercutting the monopoly. The feasibility of entry, in turn, depends on whether
the costs of entering are not recoverable ("sunk"), and how rapidly entry can occur. If the monopoly price
is not sustainable, the monopoly may engage in limit pricing, which is jargon for pricing to deter (limit)
entry.
McAfee: Introduction to Economic Analysis, http://www.introecon.com, July 24, 2006
6.241
P
PM ATC
pm ----------- -----ATC-
P2 ------------------------- -
D
q
Figure 6-22: Natural Monopoly
Historically, the United States and other nations have regulated natural monopolies like
those found in electricity, telephony and water service. An immediate problem with
regulation is that the efficient price, that is, the price that maximizes the gains from
trade, requires a subsidy from outside the industry. We see the need for a subsidy in
Figure 6-22 because the price that maximizes the gains from trade is pi, which sets the
demand (marginal value) equal to the marginal cost. At this price, however, the average
total cost exceeds the price, so that a firm with such a regulated price would lose money.
There are two alternatives. The product could be subsidized, and subsidies are used
with postal service and passenger rail in the United States, and historically for many
more products in Canada and Europe including airlines and airplane manufacture.
Alternatively, regulation could be imposed that aims to limit the price to p2, the lowest
break-even price. This is the more common strategy in the United States.
There are two strategies toward limiting the price: price-cap regulation, which directly
imposes a maximum price, and rate of return regulation, that limits the profitability of
firms. Both of these approaches induce some inefficiency of production. In both cases,
an increase in average cost may translate into additional profits for the firm, causing
regulated firms to engage in unnecessary activities.
6.5.8 Peak Load Pricing
Fluctuations in demand often require holding capacity which is used only a fraction of
the time. Hotels have off-seasons when most rooms are empty. Electric power plants
are designed to handle peak demand, usually hot summer days, with some of the
capacity standing idle on other days. Demand for trans-Atlantic airline flights is much
higher in the summer than the rest of the year. All of these examples have the similarity
that an amount of capacity - hotel space, airplane seats, electric generation capacity -
will be used over and over, which means it is used in both high demand and low demand
states. How should pricing be accomplished when demand fluctuates? This can be
McAfee: Introduction to Economic Analysis, http://www.introecon.com, July 24, 2006
6.242
thought of as a question of how to allocate the cost of capacity across several time
periods when demand systematically fluctuates.
Consider a firm that experiences two kinds of costs - a capacity cost and a marginal
cost. How should capacity be priced? This issue is applicable to a wide variety of
industries, including pipelines, airlines, telephone networks, construction, electricity,
highways, and the internet.
The basic peak-load pricing problem, pioneered by Marcel Boiteux (1922 - ), considers
two periods. The firm's profits are given by
? = p1q1 + p2q2 -fp max {q1,q2}-mc(q1+q2)-
Setting price equal to marginal costs is not sustainable, because a firm selling with price
equal to marginal cost would not earn a return on the capacity, and thus would lose
money and go out of business. Consequently, a capacity charge is necessary. The
question of peak load pricing is how the capacity charge should be allocated. This
question is not trivial because some of the capacity is used in both periods.
For the sake of simplicity, we will assume demands are independent, that is, q1 is
independent of p2 and vice versa. This assumption is often unrealistic, and generalizing
it actually doesn't complicate the problem too much. The primary complication is in
computing the social welfare when demands are functions of two prices. Independence
is a convenient starting point.
Social welfare is
W = Jpi(x)dx + Jp2(x)dx -f3 max {q1,q2} -mc(q1 +q2)
0 0
The Ramsey problem is to maximize W subject to a minimum profit condition. A
technique for accomplishing this maximization is to instead maximize
L = W+ 27.
By varying 2 , we vary the importance of profits to the maximization problem, which will
increase the profit level in the solution as 2, increases. Thus, the correct solution to the
constrained maximization problem is the outcome of the maximization of L, for some
value of &.
A useful notation is 1A, which is known as the characteristic function of the set A. This
is a function which is 1 when A is true, and zero otherwise. Using this notation, the first
order condition for the maximization of L is:
o0= = p (qq)- 011 2 - c+ qq ) +~ q, 132 -m ci
McAfee: Introduction to Economic Analysis, http://www.introecon.com, July 24, 2006
6.243
or,
pi(qi)- 11aq2 -mc 1
p1 +1E
where 1q19q2 is the characteristic function of the event ql>q2.
Similarly,
p2(q2) - p1q q2 - mc 1
p2 E2+162
Note as before that 2 -cxo yields the monopoly solution.
There are two potential types of solution. Let the demand for good 1 exceed the demand
for good 2. Either ql>q2, or the two are equal.
Case 1: q1>q2.
p1(q)-p-mc _ 1 and p2(q2)-mc 2 1
p1 +11 p2 2,1+1E2
In case 1, with all of the capacity charge allocated to good 1, quantity for good 1 still
exceeds quantity for good 2. Thus, the peak period for good 1 is an extreme peak. In
contrast, case 2 arises when assigning the capacity charge to good1 would reverse the
peak - assigning all of the capacity charge to good 1 would make period 2 the peak.
Case 2: q1=q2.
The profit equation can be written
p1(q) - mc + p2(q) - mc = 3
This equation determines q, and prices are determined from demand.
The major conclusion from peak load pricing is that either the entire cost of capacity is
allocated to the peak period, or there is no peak period in the sense that the two periods
have the same quantity demanded given the prices. That is, either the prices equalize
the quantity demanded, or the prices impose the entire cost of capacity only on one peak
period.
Moreover, the price (or, more properly, the markup over marginal cost) is proportional
to the inverse of the elasticity, which is known as Ramsey pricing.
McAfee: Introduction to Economic Analysis, http://www.introecon.com, July 24, 2006
6.244
6.6 Information
An important advantage of the price system is that it economizes on information. A
typical consumer needs to know only the prices of goods and their own personal
preferences in order to make a sensible choice of purchases, and manufacturers only
need to know the prices of goods in order to decide what to produce. Such economies of
information are an advantage over centrally-planned economies, which attempt to
direct production and consumption decisions using something other than prices, and
centrally-planned economies typically experience chronic shortages and occasional
surpluses. Shortages of important inputs to production may have dramatic effects and
the shortages aren't remedied by the price of the input rising in a centrally planned
economy, and thus often persist for long periods of time.
There are, however, circumstances where the prices are not the only necessary
information required for firms and consumers to make good decisions. In such
circumstances, information itself can lead to market failures.
6.6.1 Market for Lemons
Nobel laureate George Akerlof (1940 - ) examined the market for used cars and
considered a situation where the sellers are better informed than the buyers. This is
quite reasonable, as sellers have owned the car for a while and are likely to know its
quirks and potential problems. Akerlof showed that this differential information may
cause the used car market to collapse; that is, the information possessed by sellers of
used cars destroys the market.
To understand Akerlof's insight, suppose that the quality of used cars lies on a o to 1
scale and that the population of used cars is uniformly distributed on the interval from o
to 1. In addition, let that quality represent the value a seller places on the car, and
suppose buyers put a value that is 50% higher than the seller. Finally, the seller knows
the actual quality, while the buyer does not.
Can a buyer and seller trade in such a situation? First, note that trade is a good thing,
because the buyer values the car more than the seller. That is, both the buyer and seller
know that they should trade. But can they agree on a price? Consider a price p. At this
price, any seller who values the car less than p will be willing to trade. But because of
our uniform distribution assumption, this means the distribution of qualities of cars
offered for trade at price p will be uniform on the interval 0 to p. Consequently, the
average quality of these cars will be 1/2 p, and the buyer values these cars 50% more
which yields 3/4 p. Thus, the buyer is not willing to pay the price p for the average car
offered at price p.
The effect of the informed seller, and uninformed buyer, produces a "lemons" problem.
At any given price, all the lemons and only a few of the good cars are offered, and the
buyer - not knowing the quality of the car - isn't willing to pay as much as the actual
value of a high value car offered for sale. This causes the market to collapse; and only
the worthless cars trade at a price around zero. Economists call the differential
information an informational asymmetry.
McAfee: Introduction to Economic Analysis, http://www.introecon.com, July 24, 2006
6-245
In the real world, of course, the market has found partial or imperfect solutions to the
lemons problem identified by Akerlof. First, buyers can become informed and regularly
hire their own mechanic to inspect a car they are considering. Inspections reduce the
informational asymmetry but are costly in their own right. Second, intermediaries offer
warranties and certification to mitigate the lemons problem. The existence of both of
these solutions, which involve costs in their own right, is itself evidence that the lemons
problem is a real and significant problem, even though competitive markets find ways
to ameliorate the problems.
An important example of the lemons problem is the inventor who creates an idea that is
difficult or impossible to patent. Consider an innovation that would reduce the cost of
manufacturing computers. The inventor would like to sell it to a computer company,
but can't tell the computer company what the innovation entails prior to price
negotiations, because then the computer company could just copy the innovation.
Similarly, the computer company can't possibly offer a price for the innovation in
advance of knowing what the innovation is. As a result, such innovations usually
require the inventor to enter the computer manufacturing business, rather than selling
to an existing manufacturer, entailing many otherwise unnecessary costs.
6.6.1.1 (Exercise) In Akerlof's market for lemons model, suppose it is possible to
certify cars, verifying that they are better than a particular quality q. Thus, a
market for cars "at least as good as q" is possible. What price or prices are
possible in this market? [Hint: sellers offer cars only if q quality p.] What
quality maximizes the expected gains from trade?
6.6.2 Myerson-Satterthwaite Theorem
The lemons problem is a situation where the buyers are relatively uninformed and care
about the information held by sellers. Lemons problems are limited to situations where
the buyer isn't well-informed and can be mitigated by making information public. In
many transactions, the buyer knows the quality of the product, so lemons concerns
aren't a significant issue. There can still be a market failure, however, if there are a
limited number of buyers and sellers.
Consider the case of one buyer and one seller bargaining over the sale of a good. The
buyer knows his own value v for the good, but not the seller's cost. The seller knows her
own cost c for the good, but not the buyer's value. The buyer views the seller's cost as
uniformly distributed on the interval [o,1], and similarly the seller views the buyer's
value as uniformly distributed on [0,1].80 Can efficient trade take place? Efficient trade
requires that trade occurs whenever v>c, and the remarkable answer is that it is
impossible to arrange efficient trade if the buyer and seller are to trade voluntarily. This
8o The remarkable fact proved by Roger Myerson and Mark Satterthwaite (Efficient Mechanisms for
Bilateral Trade, Journal of Economic Theory, 28, 1983, 265-281) is that the distributions don't matter;
the failure of efficient trade is a fully general property. Philip Reny and Preston McAfee (Correlated
Information and Mechanism Design, Econometrica 6o, No. 2, March 1992, 395-421) show the nature of
the distribution of information matters, and Preston McAfee (Efficient Allocation with Continuous
Quantities, Journal of Economic Theory 53, no. 1, February 1991: 51-74.) showed that continuous
quantities can overturn the Myerson-Sattertihwaite theorem.
McAfee: Introduction to Economic Analysis, http://www.introecon.com, July 24, 2006
6.246
is true even if a third party is used to help arrange trade, provided the third party
doesn't subsidize the transaction.
The total gains from trade under efficiency are
10 1 2
V 1 'dv =-.
v-cdcdv = 2 6
00 0
A means of arranging trade, or a mechanism, asks the buyer and seller for their value
and cost, respectively, and then orders trade if the value exceeds the cost, and dictates a
payment p by the buyer to the seller. Buyers need not make honest reports to the
mechanism, however, and the mechanisms must be designed to induce the buyer and
seller to report honestly to the mechanism, so that efficient trades can be arranged.81
Consider a buyer who actually has value v but reports a value r. The buyer trades with
the seller if the seller has a cost less than r, which occurs with probability r.
u(r,v) = yr - Ecp(r, c).
The buyer gets the actual value v with probability r, and makes a payment that depends
on the buyer's report and the seller's report, but we can take expectations over the
seller's report to eliminate it (from the buyer's perspective), and this is denoted Ecp(r,
c), which is just the expected payment given the report r. In order for the buyer to
choose to be honest, u must be maximized at r=v for every v, for otherwise some buyers
would lie and some trades would not be efficiently arranged. Thus, we can conclude82
d
u(v,v)=u1(v,v)+u2(v,v)=u2(v,v)=r =v.
dv r=v
The first equality is just the total derivative of u(v,v), because there are two terms; the
second equality because u is maximized over the first argument r at r=v, and the first
order condition insures ui = o. Finally, u2 is just r, and we are evaluating the derivative
at the point r = v. A buyer who has a value v + A, but who reports v, trades with
probability v and makes the payment Ecp(v, c). Such a buyer gets Av more in utility
than the buyer with value v. Thus a A increase in value produces an increase in utility of
d
at least Av, showing that u(v + A, v + A) u(v, v) + Av and hence that u(v, v) v. A
dv
similar argument considering a buyer with value v who reports v + A shows that equality
occurs.
81 Inducing honesty is without loss of generality. Suppose that the buyer of type v reported the type z(v).
Then we can add a stage to the mechanism, where the buyer reports a type, which is converted via the
function z to a report, and then that report given to the original mechanism. In the new mechanism,
reporting v is tantamount to reporting z(v) to the original mechanism.
82 We maintain an earlier notation that the subscript refers to a partial derivative, so that if we have a
functionf,fi is the partial derivative off with respect to the first argument off.
McAfee: Introduction to Economic Analysis, http://www.introecon.com, July 24, 2006
6.247
The value u(v,v) is the gain accruing to a buyer with value v who reports having value v.
Since the buyer with value o gets zero, the total gain accruing to the average buyer can
be computed by integrating by parts
1 11
Svu(v,v)dv = -(1-uu(v,v) + J(1-v)y du =f(1-v)vdv =-.
J-u-v vdv -'- d
O V=0O 0 0
In the integration by parts, dv = d -(1-v) is used. The remarkable conclusion is that, if
the buyer is induced to truthfully reveal the buyer's value, the buyer must obtain the
entire gains from trade! This is actually a quite general proposition. If you offer the
entire gains from trade to a party, they are induced to maximize the gains from trade.
Otherwise, they will want to distort away from maximizing the entire gains from trade,
which will result in a failure of efficiency.
The logic with respect to the seller is analogous: the only way to get the seller to report
her cost honestly is to offer her the entire gains from trade.
6.6.2.1 (Exercise) Let h(r, c) be the gains of a seller who has cost c and reports r,
h(r, c) = p(v, r) - (1-r)c.
Noting that the highest cost seller (c=i) never sells and thus obtains zero profits,
show that honesty by the seller implies the expected value of h isY6.
The Myerson-Satterthwaite theorem shows that the gains from trade are insufficient to
induce honesty by both parties. (Indeed, they are half the necessary amount!) Thus,
any mechanism for arranging trades between the buyer and the seller must suffer some
inefficiency. Generally this occurs because buyers act like they value the good less than
they do, and sellers act like their costs are higher than they truly are.
It turns out that the worst case scenario is a single buyer and a single seller. As markets
get "thick," the per capita losses converge to zero, and markets become efficient. Thus,
informational problems of this kind are a "small numbers" issue. However, many
markets do in fact have small numbers of buyers or sellers. In such markets, it seems
likely that informational problems will be an impediment to efficient trade.
6.6.3 Signaling
An interesting approach to solving informational problems involves signaling.8 3
Signaling, in economic jargon, means expenditures of time or money whose purpose is
to convince others of something. Thus, people signal wealth by wearing Rolex watches,
driving expensive cars or sailing in the America's Cup. They signal erudition by tossing
out quotes from Kafka or Tacitus into conversations. They signal being chic by wearing
the right clothes and listening to cool music. Signaling is also rampant in the animal
83 Signaling was introduced by Nobel laureate Michael Spence in his dissertation, part of which was
reprinted in "Job Market Signaling," Quarterly Journal of Economics 87, August 1973, 355-74.
McAfee: Introduction to Economic Analysis, http://www.introecon.com, July 24, 2006
6.248
world, from peacock feathers to elk battles and the subject of a vibrant and related
research program.
A university education serves not just to educate, but also to signal the ability to learn.
Businesses often desire employees who are able to adapt to changing circumstances, and
who can easily and readily learn new strategies and approaches. Education signals such
abilities because it will easier for quick learners to perform well in university. A simple
model suffices to illustrate the point. Suppose there are two types of people. Type A has
a low cost cA of learning, and type B has a higher cost cB of learning. It is difficult to
determine from an interview whether someone is type A or not. Type A is worth more to
businesses, and the competitive wage wA (expressed as a present value of lifetime
earnings) for type A's is higher than the wage wB for type B's.
A person can signal that they are a type A by taking a sufficient amount of education.
Suppose the person devotes an amount of time x to learning in university, thus incurring
the cost cA x. If x is large enough so that
WA - CA X> WB> WA - CB X,
it pays the type A to obtain the education, but not the type B, if education in fact signals
that the student is type A. Thus, a level of education x in this case signals a trait (ease of
learning) that is valued by business, and it does so by voluntary choice - those with a
high cost of learning choose not to obtain the education, even though they could do it.
This works as a signal because only type A would voluntarily obtain the education in
return for being perceived to be a type A.
There are several interesting aspects to this kind of signaling. First, the education
embodied in x need not be valuable in itself; the student could be studying astronomy or
ancient Greek, neither of which are very useful in most businesses, but are nevertheless
strong signals of the ability to learn. Second, the best subject matter for signaling is that
in which the difference in cost between the type desired by employers and the less
desirable type is greatest, that is, where cB - cA IS greatest. Practical knowledge is
somewhat unlikely to make this difference great; instead, challenging abstract problem-
solving may be a better separator. Clearly, it is desirable to have the subject matter be
useful, if it can still do the signaling job. But interpreting long medieval poems could
more readily signal the kind of flexible mind desired in management than studying
accounting, not because the desirable type is good at it, or that it is useful, but because
the less desirable type is so much worse at it.
Third, one interprets signals by asking "what kinds of people would make this choice?"
while understanding that the person makes the choice hoping to send the signal.
Successful law firms have very fine offices, generally much finer than the offices of their
clients. Moreover, there are back rooms at most law firms, where much of the real work
is done, that aren't nearly so opulent. The purpose of the expensive offices is to signal
success, essentially making the statement that
"we couldn't afford to waste money on such expensive offices if we weren't very
successful. Thus, you should believe we are successful."
McAfee: Introduction to Economic Analysis, http://www.introecon.com, July 24, 2006
6.249
The law firm example is similar to the education example. Here, the cost of the
expenditures on fancy offices is different for different law firms because more successful
firms earn more money and thus value the marginal dollar less. Consequently, more
successful firms have a lower cost of a given level of office luxury. What is interesting
about signaling is that it is potentially quite wasteful. A student spends four years
studying boring poems and dead languages in order to demonstrate a love of learning,
and a law firm pays $75,000 for a conference table that it rarely uses and gets no
pleasure out of, in order to convince a client that the firm is extremely successful. In
both cases, it seems like a less costly solution should be available. The student can take
standardized tests, and the law firm could show its win-loss record to the potential
client. But standardized tests may measure test-taking skills rather than learning
ability, especially if what matters is the learning ability over a long time horizon. Win-
loss records can be "massaged," and in the majority of all legal disputes, the case settles
and both sides consider themselves "the winner." Consequently, statistics may not be a
good indicator of success, and the expensive conference table a better guide.
McAfee: Introduction to Economic Analysis, http://www.introecon.com, July 24, 2006
6-250
7 Strategic Behavior
Competitive theory studies price-taking consumers and firms, that is, people who can't
individually affect the transaction prices. The assumption that market participants take
prices as given is justified only when there are many competing participants. We have
also examined monopoly, precisely because a monopoly by definition doesn't have to
worry about competitors. Strategic behavior involves the examination of the
intermediate case, where there are few enough participants that they take each other
into account and their actions individually matter, and where the behavior of any one
participant influences choices of the other participants. That is, participants are
strategic in their choice of action, recognizing that their choice will affect choices made
by others.
The right tool for the job of examining strategic behavior in economic circumstances is
game theory, the study of how people play games. Game theory was pioneered by the
mathematical genius John von Neumann (1903-1957). Game theory has also been very
influential in the study of military strategy, and indeed the strategy of the cold war
between the United States and the U.S.S.R. was guided by game theoretic analyses.84
7.1 Games
The theory of games provides a description of games that fits common games like poker
or the board game "Monopoly" but will cover many other situations as well. In any
game, there is a list of players. Games generally unfold over time; at each moment in
time, players have information, possibly incomplete, about the current state of play, and
a set of actions they can take. Both information and actions may depend on the history
of the game prior to that moment. Finally, players have payoffs, and are assumed to
play in such a way as to maximize their expected payoff, taking into account their
expectations for the play of others. When the players, their information and available
actions, and payoffs have been specified, we have a game.
7.1.1 Matrix Games
The simplest game is called a matrix payoff game with two players. In a matrix payoff
game, all actions are chosen simultaneously. It is conventional to describe a matrix
payoff game as played by a row player and a column player. The row player chooses a
row in a matrix; the column player simultaneously chooses a column. The outcome of
the game is a pair of payoffs where the first entry is the payoff of the row player and the
second is the payoff of the column player. Table 7-1 provides an example of a "2 x 2"
matrix payoff game, the most famous game of all, which is known as the prisoner's
dilemma.
84 An important reference for game theory is John von Neumann (1903-1957) and Oskar Morgenstern
(1902-1977), Theory of Games and Economic Behavior, Princeton: Princeton University Press, 1944.
Important extensions were introduced by John Nash (1928 - ), the mathematician made famous by Sylvia
Nasar's delightful book A Beautiful Mind (Simon & Schuster, 1998). Finally, applications in the military
arena were pioneered by Nobel Laureate Thomas Schelling (1921 - ), The Strategy of Conflict,
Cambridge: Cambridge University Press, 1960.
McAfee: Introduction to Economic Analysis, http://www.introecon.com, July 24, 2006
7-251
Table 7-1: The Prisoner's Dilemma
Column
Confess Don't
3 Confess (-10,-10) (0,-20)
2 Don't (-20,0) (-1,-1)
In the prisoner's dilemma, two criminals named Row and Column have been
apprehended by the police and are being questioned separately. They are jointly guilty
of the crime. Each player can choose either to confess or not. If Row confesses, we are
in the top row of the matrix (corresponding to the row labeled Confess). Similarly, if
Column confesses, the payoff will be in the relevant column. In this case, if only one
player confesses, that player goes free and the other serves twenty years in jail. (The
entries correspond to the number of years lost to prison. The first entry is always Row's
payoff, the second Column's payoff.) Thus, for example, if Column confesses and Row
does not, the relevant payoff is the first column and the second row, in reverse color in
Table 7-2.
Table 7-2: Solving the Prisoner's Dilemma
0
Column
Confess Don't
Confess (-1o,-1o) (0,-20)
Don't (-,i
If Column confesses and Row does not, Row loses twenty years, and Column loses no
years, that is, goes free. This is the payoff (-20,0) in reverse color in Table 7-2. If both
confess, they are both convicted and neither goes free, but they only serve ten years
each. Finally, if neither confesses, there is a ten percent chance they are convicted
anyway (using evidence other than the confession), in which case they average a year
lost each.
The prisoner's dilemma is famous partly because it is readily solvable. First, Row has a
strict advantage to confessing, no matter what Column is going to do. If Column
confesses, Row gets -10 from confessing, -20 from not, and thus is better off from
confessing. Similarly, if Column doesn't confess, Row gets o from confessing, -1 from
not confessing, and is better off confessing. Either way, no matter what Column does,
Row should choose to confess.85 This is called a dominant strategy, a strategy that is
optimal no matter what the other players do.
The logic is exactly similar for Column: no matter what Row does, Column should
choose to confess. That is, Column also has a dominant strategy, to confess. To
establish this, first consider what Column's best action is, when Column thinks Row will
confess. Then consider Column's best action when Column thinks Row won't confess.
85 If Row and Column are friends are care about each other, that should be included as part of the payoffs.
Here, there is no honor or friendship among thieves, and Row and Column only care about what they
themselves will get.
McAfee: Introduction to Economic Analysis, http://www.introecon.com, July 24, 2006
7-252
Either way, Column gets a higher payoff (lower number of years lost to prison) by
confessing.
The presence of a dominant strategy makes the prisoner's dilemma particularly easy to
solve. Both players should confess. Note that this gets them ten years each in prison,
and thus isn't a very good outcome from their perspective, but there is nothing they can
do about it in the context of the game, because for each, the alternative to serving ten
years is to serve twenty years. This outcome is referred to as (Confess, Confess), where
the first entry is the row player's choice, and the second entry is the column player's
choice.
Consider an entry game, played by Microsoft (the row player) and Piuny (the column
player), a small start-up company. Both Microsoft and Piuny are considering entering a
new market for an online service. The payoff structure is
Table 7-3: An Entry Game
Piuny
Enter Don't
Enter (2,-2) (5,0)
2 Don't (0,5) (o,o)
In this case, if both companies enter, Microsoft ultimately wins the market, and earns 2,
and Piuny loses 2. If either firm has the market to itself, they get 5 and the other firm
gets zero. If neither enters, both get zero. Microsoft has a dominant strategy to enter: it
gets 2 when Piuny enters, 5 when Piuny doesn't, and in both cases does better than
when Microsoft doesn't enter. In contrast, Piuny does not have a dominant strategy:
Piuny wants to enter when Microsoft doesn't, and vice-versa. That is, Piuny's optimal
strategy depends on Microsoft's action, or, more accurately, Piuny's optimal strategy
depends on what Piuny believes Microsoft will do.
Piuny can understand Microsoft's dominant strategy, if it knows the payoffs of
Microsoft.86 Thus, Piuny can conclude that Microsoft is going to enter, and this means
that Piuny should not enter. Thus, the equilibrium of the game is for MS to enter and
Piuny not to enter. This equilibrium is arrived at by the iterated elimination of
dominated strategies, which sounds like jargon but is actually plain speaking. First, we
eliminated Microsoft's dominated strategy in favor of its dominant strategy. Microsoft
had a dominant strategy to enter, which means the strategy of not entering is dominated
by the strategy of entering, so we eliminated the dominated strategy. That leaves a
simplified game in which Microsoft enters:
86 It isn't so obvious that one player will know the payoffs of another player, and that often causes players
to try to signal that they are going to play a certain way, that is, to demonstrate commitment to a
particular advantageous strategy. Such topics are taken up in business strategy and managerial
economics.
McAfee: Introduction to Economic Analysis, http://www.introecon.com, July 24, 2006
7-253
Table 7-4; Eliminating a Dominated Strategy
Piuny
Enter Don't
Enter (2,-2)
(5,0)
I I I I
In this simplified game, after the elimination of Microsoft's dominated strategy, Piuny
also has a dominant strategy: not to enter. Thus, we iterate and eliminate dominated
strategies again, this time eliminating Piuny's dominated strategies, and wind up with a
single outcome: Microsoft enters, and Piuny doesn't. The iterated elimination of
dominated strategies solves the game.87
Here is another game, with three strategies for each player.
Table 7-5: A 3 X 3 Game
0O
Column
Left Center Right
Top (-5,-1) (2,2) (3,3)
Middle (1,-3) (1,2) (1,1)
Bottom (o,10) (o,o) (0,-10)
The process of iterated elimination of dominated strategies is illustrated by actually
eliminating the rows and columns, as follows. A reverse color (white writing on black
background) indicates a dominated strategy.
Middle dominates bottom for Row, yielding:
Table 7-6: Eliminating a Dominated Strategy
Column
Left Center Right
Top (-5,-1) (2,2) (3,3)
0O
Middle (1,-3)
(1,2) (i,i)
Bottom (o,-) (o,o) (,-10)
With bottom eliminated, Left is now dominated for Column by either Center or Right,
which eliminates the left column.
87 A strategy may be dominated not by any particular alternate strategy but by a randomization over other
strategies, which is an advanced topic not considered here.
McAfee: Introduction to Economic Analysis, http://www.introecon.com, July 24, 2006
7-254
Table 7-7: Eliminating Another Dominated Strategy
Column
3
Center Ri ht
To (2,2) (3,3)
Middle (1,2) (i,i)
With Left and Bottom eliminated, Top now dominates Middle for Row.
Table 7-8: Eliminating a Third Dominated Strategy
Column
Left Center Right
To (2,2) (3,3)
Middle (1,-3) (1,2) (1,1)
Bottom (o,10) (OAo (0,-10)
Finally, Column chooses Right over Center, yielding a unique outcome after the iterated
elimination of dominated strategies, which is (Top, Right).
Table 7-9: Game Solved
Column
0O
Left CenterRih
Top (-5,-1) (2,2)
Middle (1,-3) (1,2) (1,1)
Bottom (o,10) (o,o) (0,-10)
The iterated elimination of dominated strategies is a useful concept, and when it applies,
the predicted outcome is usually quite reasonable. Certainly it has the property that no
player has an incentive to change their behavior given the behavior of others. However,
there are games where it doesn't apply, and these games require the machinery of a
Nash equilibrium, named for Nobel laureate John Nash (1928 - ).
7.1.2 Nash Equilibrium
In a Nash equilibrium, each player chooses the strategy that maximizes their expected
payoff, given the strategies employed by others. For matrix payoff games with two
players, a Nash equilibrium requires that the row chosen maximizes the row player's
payoff, given the column chosen by the column player, and the column, in turn,
maximizes the column player's payoff given the row selected by the row player. Let us
consider first the prisoner's dilemma, which we have already seen.
McAfee: Introduction to Economic Analysis, http://www.introecon.com, July 24, 2006
7-255
Table 7-1o: Prisoner's Dilemma Again
Column
Confess Don't
: Confess (-10,-10) (0,-20)
e Don't (-20,0) (-1,-1)
Given that the row player has chosen to confess, the column player also chooses
confession because -10 is better than -20. Similarly, given that the column player
chooses confession, the row player chooses confession, because -10 is better than -20.
Thus, for both players to confess is a Nash equilibrium. Now let us consider whether
any other outcome is a Nash equilibrium. In any outcome, at least one player is not
confessing. But that player could get a higher payoff by confessing, so no other outcome
could be a Nash equilibrium.
The logic of dominated strategies extends to Nash equilibrium, except possibly for ties.
That is, if a strategy is strictly dominated, it can't be part of a Nash equilibrium. On the
other hand, if it involves a tied value, a strategy may be dominated but still part of a
Nash equilibrium.
The Nash equilibrium is justified as a solution concept for games as follows. First, if the
players are playing a Nash equilibrium, no one has an incentive to change their play or
re-think their strategy. Thus, the Nash equilibrium has a "steady state" aspect in that no
one wants to change their own strategy given the play of others. Second, other potential
outcomes don't have that property: if an outcome is not a Nash equilibrium, then at
least one player does have an incentive to change what they are doing. Outcomes that
aren't Nash equilibria involve mistakes for at least one player. Thus, sophisticated,
intelligent players may be able to deduce each other's play, and play a Nash equilibrium
Do people actually play Nash equilibria? This is a controversial topic and mostly beyond
the scope of this book, but we'll consider two well-known games: Tic-Tac-Toe (see, e.g.
http://www.mcafee.cc/Bin/tictactoe/index.html) and Chess. Tic-Tac-Toe is a relatively
simple game, and the equilibrium is a tie. This equilibrium arises because each player
has a strategy that prevents the other player from winning, so the outcome is a tie.
Young children play Tic-Tac-Toe and eventually learn how to play equilibrium
strategies, at which point the game ceases to be very interesting since it just repeats the
same outcome. In contrast, it is known that Chess has an equilibrium, but no one knows
what it is. Thus, at this point we don't know if the first mover (White) always wins, or
the second mover (Black) always wins, or if the outcome is a draw (neither is able to
win). Chess is complicated because a strategy must specify what actions to take given
the history of actions, and there are a very large number of potential histories of the
game thirty or forty moves after the start. So we can be quite confident that people are
not (yet) playing Nash equilibria to the game of Chess.
The second most famous game in game theory is the battle of the sexes. The battle of
the sexes involves a married couple who are going to meet each other after work, but
haven't decided where they are meeting. Their options are a baseball game or the ballet.
McAfee: Introduction to Economic Analysis, http://www.introecon.com, July 24, 2006
7-256
Both prefer to be with each other, but the man prefers the baseball game and the woman
prefers the ballet. This gives payoffs something like this:
Table 7-11: The Battle of the Sexes
Woman
Baseball Ballet
r Baseball (3,2) (1,1)
2 Ballet (o,o) (2,3)
The man would rather that they both go to the baseball game, and the woman that they
both go to the ballet. They each get 2 payoff points for being with each other, and an
additional point for being at their preferred entertainment. In this game, iterated
elimination of dominated strategies eliminates nothing. You can readily verify that
there are two Nash equilibria: one in which they both go to the baseball game, and one
in which they both go to ballet. The logic is: if the man is going to the baseball game, the
woman prefers the 2 points she gets at the baseball game to the single point she would
get at the ballet. Similarly, if the woman is going to the baseball game, the man gets
three points going there, versus zero at the ballet. Thus, for both to go to the baseball
game is a Nash equilibrium. It is straightforward to show that for both to go to the
ballet is also a Nash equilibrium, and finally that neither of the other two possibilities,
involving not going to the same place, is an equilibrium.
Now consider the game of matching pennies. In this game, both the row player and the
column player choose heads or tails, and if they match, the row player gets the coins,
while if they don't match, the column player gets the coins. The payoffs are provided in
the next table.
Table 7-12: Matching Pennies
0
I--
Column
Heads Tails
Heads (1,-1) (-1,1)
Tails (-,) ( -1
You can readily verify that none of the four possibilities represents a Nash equilibrium.
Any of the four involves one player getting -1; that player can convert -1 to 1 by changing
his or her strategy. Thus, whatever the hypothesized equilibrium, one player can do
strictly better, contradicting the hypothesis of a Nash equilibrium. In this game, as
every child who plays it knows, it pays to be unpredictable, and consequently players
need to randomize. Random strategies are known as mixed strategies, because the
players mix across various actions.
7.1.3 Mixed Strategies
Let us consider the matching pennies game again.
McAfee: Introduction to Economic Analysis, http://www.introecon.com, July 24, 2006
7-257
Table 7-13: Matching Pennies Again
Column
Heads Tails
i Heads (i,-i) (-i,i)
S Tails (-1,1) (1,-1)
Suppose that Row believes Column plays Heads with probability p. Then if Row plays
Heads, Row gets 1 with probability p and -1 with probability (i-p), for an expected value
of 2p - 1. Similarly, if Row plays Tails, Row gets -1 with probability p (when Column
plays Heads), and 1 with probability (1-p), for an expected value of 1 - 2p. This is
summarized in the next table.
Table 7-14: Mixed Strategy in Matching Pennies
Column
Heads Tails
i Heads (i,-i) (-1,1) 1 + -1(1-p)=2p-1
N Tails (-1,1) (1,-1) -1p + 1(1-p)=1-2p
If 2p - 1 > 1 - 2p, then Row is better off on average playing Heads than Tails. Similarly,
if 2p - 1 < 1 - 2p, Row is better off playing Tails than Heads. If, on the other hand, 2p -
1 = 1 - 2p, then Row gets the same payoff no matter what Row does. In this case Row
could play Heads, could play Tails, or could flip a coin and randomize Row's play.
A mixed strategy Nash equilibrium involves at least one player playing a randomized
strategy, and no player being able to increase their expected payoff by playing an
alternate strategy. A Nash equilibrium without randomization is called a pure strategy
Nash equilibrium.
Note that that randomization requires equality of expected payoffs. If a player is
supposed to randomize over strategy A or strategy B, then both of these strategies must
produce the same expected payoff. Otherwise, the player would prefer one of them, and
wouldn't play the other.
Computing a mixed strategy has one element that often appears confusing. Suppose
Row is going to randomize. Then Row's payoffs must be equal, for all strategies Row
plays with positive probability. But that equality in Row's payoffs doesn't determine the
probabilities with which Row plays the various rows. Instead, that equality in Row's
payoffs will determine the probabilities with which Column plays the various columns.
The reason is that it is Column's probabilities that determine the expected payoff for
Row; if Row is going to randomize, then Column's probabilities must be such that Row
is willing to randomize.
Thus, for example, we computed the payoff to Row of playing Heads, which was 2p - 1,
where p was the probability Column played Heads. Similarly, the payoff to Row of
playing Tails was 1 - 2p. Row is willing to randomize if these are equal, which solves for
p = W/.
McAfee: Introduction to Economic Analysis, http://www.introecon.com, July 24, 2006
7-258
7.1.3.1 (Exercise) Let q be the probability that Row plays Heads. Show that Column is
willing to randomize if, and only if, q = 1/2. (Hint: First compute Column's
expected payoff when Column plays Heads, and then Column's expected payoff
when Column plays Tails. These must be equal for Column to randomize.)
Now let's try a somewhat more challenging example, and revisit the battle of the sexes.
Table 7-15: Mixed Strategy in Battle of the Sexes
Woman
Baseball Ballet
Baseball (3,2) (i,i)
2 Ballet (o,o) (2,3)
This game has two pure strategy Nash equilibria: (Baseball,Baseball) and (Ballet,Ballet).
Is there a mixed strategy? To compute a mixed strategy, let the Woman go to the
baseball game with probability p, and the Man go to the baseball game with probability
q. Table 7-16 contains the computation of the mixed strategy payoffs for each player.
Table 7-16: Full Computation of the Mixed Strategy
Woman
Baseball (p) Ballet (i-p) Man's E Payoff
c Baseball (prob ) (3,2) (i,i) 3 + 1(1-p)=1+2
Ballet (prob 1-q) (o,o) (2,3) O + 2(1-p)=2-2p
Woman's E Payoff 2q + o(1-q)=2q ig + 3(1-q)=3-2q
For example, if the Man (row player) goes to the baseball game, he gets 3 when the
Woman goes to the baseball game (probability p) and otherwise gets 1, for an expected
payoff of 3p + i(i-p) = 1 + 2p. The other calculations are similar but you should
definitely run through the logic and verify each calculation.
A mixed strategy in the Battle of the Sexes game requires both parties to randomize
(since a pure strategy by either party prevents randomization by the other). The Man's
indifference between going to the baseball game and the ballet requires 1+2p = 2 - 2p,
which yields p = 1/4. That is, the Man will be willing to randomize which event he
attends if the Woman is going to the ballet 3/4 of the time, and otherwise to the baseball
game. This makes the Man indifferent between the two events, because he prefers to be
with the Woman, but he also likes to be at the baseball game; to make up for the
advantage that the game holds for him, the woman has to be at the ballet more often.
Similarly, in order for the Woman to randomize, the Woman must get equal payoffs
from going to the game and going to the ballet, which requires 2q = 3 - 2q, or q = %/4.
Thus, the probability that the Man goes to the game is 3/, and he goes to the ballet 1/ of
the time. These are independent probabilities, so to get the probability that both go to
McAfee: Introduction to Economic Analysis, http://www.introecon.com, July 24, 2006
7-259
the game, we multiply the probabilities, which yields 6. The next table fills in the
probabilities for all four possible outcomes.
Table 7-17: Mixed Strategy Probabilities
Woman
Baseball Ballet
Baseball 3/ 9
Ballet 1/
/16 i6
Note that more than half the time, (Baseball, Ballet) is the outcome of the mixed
strategy, and the two people are not together. This lack of coordination is a feature of
mixed strategy equilibria generally. The expected payoffs for both players are readily
computed as well. The Man's payoff was 1+2p = 2 - 2p, and since p = 1/4, the Man
obtained 1 1/2. A similar calculation shows the Woman's payoff is the same. Thus, both
do worse than coordinating on their less preferred outcome. But this mixed strategy
Nash equilibrium, undesirable as it may seem, is a Nash equilibrium in the sense that
neither party can improve their payoff, given the behavior of the other party.
In the Battle of the sexes, the mixed strategy Nash equilibrium may seem unlikely, and
we might expect the couple to coordinate more effectively. Indeed, a simple call on the
telephone should rule out the mixed strategy. So let's consider another game related to
the Battle of the Sexes, where a failure of coordination makes more sense. This is the
game of "Chicken." Chicken is played by two drivers driving toward each other, trying to
convince the other to yield, which involves swerving into a ditch. If both swerve into the
ditch, we'll call the outcome a draw and both get zero. If one swerves and the other
doesn't, the swerver loses and the other wins, and we'll give the winner one point.88 The
only remaining question is what happens when both don't yield, in which case a crash
results. In this version, that has been set at four times the loss of swerving, but you can
change the game and see what happens.
Table 7-18: Chicken
Column
0:
Swerve Don't
Swerve (o,o) (-1,1)
Don't (1,-1) (-4,-4)
This game has two pure strategy equilibria: (Swerve, Don't) and (Don't, Swerve). In
addition, it has a mixed strategy. Suppose Column swerves with probability p. Then
88 Note that adding a constant to a player's payoffs, or multiplying that player's payoffs by a positive
constant, doesn't affect the Nash equilibria, pure or mixed. Therefore, we can always let one outcome for
each player be zero, and another outcome be one.
McAfee: Introduction to Economic Analysis, http://www.introecon.com, July 24, 2006
7.260
Row gets op + -1(1-p) from swerving, 1p + (-4)(1-p) from not swerving, and Row will
randomize if these are equal, which requires p = 3/4. That is, the probability that
Column swerves, in a mixed strategy equilibrium is 3/4. You can verify that the Row
player has the same probability by setting the probability that Row swerves equal to q
and computing Column's expected payoffs. Thus, the probability of a collision is 16 in
the mixed strategy equilibrium.
The mixed strategy equilibrium is more likely in some sense in this game; if the players
already knew which player would yield, they wouldn't actually need to play the game.
The whole point of the game is to find out who will yield, which means it isn't known in
advance, which means the mixed strategy equilibrium is in some sense the more
reasonable equilibrium.
Paper, Scissors, Rock is a child's game in which two children simultaneously choose
paper (hand held flat), scissors (hand with two fingers protruding to look like scissors)
or rock (hand in a fist). The nature of the payoffs is that paper beats rock, rock beats
scissors, and scissors beat paper. This game has the structure
Table 7-19: Paper, Scissors, Rock
0O
Column
Paper Scissors Rock
Paper (o,o) (-1,1) (1,-1)
Scissors (i,-i) (o,o) (-i,i)
Rock (-i,i) (i,-i) (o,o)
7.1.3.2 (Exercise) Show that, in the Paper, Scissors, Rock game, there are no pure
strategy equilibria. Show that playing all three actions with equal likelihood is a
mixed strategy equilibrium.
7.1.3.3 (Exercise) Find all equilibria of the following games:
1
O
2
O
Column
Left Right
Up (3,2) (11,1)
Down (4,5) (8,o)
Column
Left Right
Up (3,3) (o,o)
Down (4,5) (8,o)
McAfee: Introduction to Economic Analysis, http://www.introecon.com, July 24, 2006
7.261
3
0
,
Up
Left
(0,3
(4,0
Column
Right
) (3,0)
) (0,4)
---I
I I I I
4
,I
I
0
I--
Up
Down
Left
(7,2
(8,7
Column
Right
(O,9)
(8,8)
----I
I I I I
5
,
C
6
0
Up
Left
(1,1)
Column
Right
(2,4)
---I
Down (4,1) (3,2)
Column
Left Right
Up (4,2) (2,3)
Down (3,8) (1,5)
7.1.4 Examples
Our first example concerns public goods. In this game, each player can either
contribute, or not. For example, two roommates can either clean their apartment, or
not. If they both clean, the apartment is nice. If one cleans, that roommate does all the
work and the other gets half of the benefits. Finally, if neither clean, neither is very
happy. This suggests payoffs like:
Table 7-20: Cleaning the Apartment
Column
Clean Don't
Clean (io,io) (0,15)
Don't (15,o) (2,2)
0
I--
You can verify that this game is similar to the prisoner's dilemma, in that the only Nash
equilibrium is the pure strategy in which neither player cleans. This is a game theoretic
version of the tragedy of the commons - even though the roommates would both be
better off if both cleaned, neither do. As a practical matter, roommates do solve this
problem, using strategies that we will investigate when we consider dynamic games.
McAfee: Introduction to Economic Analysis, http://www.introecon.com, July 24, 2006
7.262
Table 7-21: Driving on the Right
Column
Left Right
: Left (1,1) (o,o)
P Right (o,o) (1,1)
The important thing about the side of the road the cars drive on is not that it is the right
side but that it is the same side. This is captured in the Driving on the Right game
above. If both players drive on the same side, then they both get one point, otherwise
they get zero. You can readily verify that there are two pure strategy equilibria,
(Left,Left) and (Right,Right), and a mixed strategy equilibrium with equal probabilities.
Is the mixed strategy reasonable? With automobiles, there is little randomization. On
the other hand, people walking down hallways often seem to randomize whether they
pass on the left or the right, and sometimes do that little dance where they try to get past
each other, one going left and the other going right, then both simultaneously reversing,
unable to get out of each other's way. That dance suggests that the mixed strategy
equilibrium is not as unreasonable as it seems in the automobile application.89
Table 7-22: Bank Location Game
NYC
No Concession Tax Rebate
No Concession (30,10) (10,20)
-A
Tax Rebate (20,10)
(20,0)
Consider a foreign bank that is looking to open a main office and a smaller office in the
United States. The bank narrows its choice for main office to either New York (NYC) or
Los Angeles (LA), and is leaning toward Los Angeles. If neither city does anything, LA
will get $30 million in tax revenue and New York ten million. New York, however, could
offer a $10 million rebate, which would swing the main office to New York, but now New
York would only get a net of $20 M. The discussions are carried on privately with the
bank. LA could also offer the concession, which would bring the bank back to LA.
89 Continental Europe drove on the left until about the time of the French revolution. At that time, some
individuals began driving on the right as a challenge to royalty who were on the left, essentially playing
the game of chicken with the ruling class. Driving on the right became a symbol of disrespect for royalty.
The challengers won out, forcing a shift to driving on the right. Besides which side one drives on, another
coordination game involves whether one stops or goes on red. In some locales, the tendency for a few
extra cars to go as a light changes from green to yellow to red forces those whose light changes to green to
wait, and such a progression can lead to the opposite equilibrium, where one goes on red and stops on
green. Under Mao Tse-tung, the Chinese considered changing the equilibrium to going on red and
stopping on green (because 'red is the color of progress') but wiser heads prevailed and the plan was
scrapped.
McAfee: Introduction to Economic Analysis, http://www.introecon.com, July 24, 2006
7.263
7.1.4.1 (Exercise) Verify that the bank location game has no pure strategy equilibria,
and that there is a mixed strategy equilibrium where each city offers a rebate
with probability 1/2.
Table 7-23: Political Mudslinging
Republican
Clean Mud
S Clean (3,1) (1,2)
Q Mud (2,1) (2,0)
On the night before the election, a Democrat is leading the Wisconsin senatorial race.
Absent any new developments, the Democrat will win, and the Republican will lose.
This is worth 3 to the Democrat, and the Republican, who loses honorably, values this
outcome at one. The Republican could decide to run a series of negative advertisements
("throwing mud") against the Democrat, and if so, the Republican wins although loses
his honor, which he values at 1, and so only gets 2. If the Democrat runs negative ads,
again the Democrat wins, but loses his honor, so only gets 2. These outcomes are
represented in the Mudslinging game above.
7.1.4.2 (Exercise) Show that the only Nash equilibrium is a mixed strategy with equal
probabilities of throwing mud and not throwing mud.
7.1.4.3 (Exercise) Suppose that voters partially forgive a candidate for throwing mud
when the rival throws mud, so that the (Mud, Mud) outcome has payoff (2.5,.5).
How does the equilibrium change?
You have probably had the experience of trying to avoid encountering someone, who we
will call Rocky. In this instance, Rocky is actually trying to find you. The situation is
that it is Saturday night and you are choosing which party, of two possible parties, to
attend. You like party 1 better, and if Rocky goes to the other party, you get 20. If Rocky
attends party 1, you are going to be uncomfortable and get 5. Similarly, Party 2 is worth
15, unless Rocky attends, in which case it is worth o. Rocky likes Party 2 better (these
different preferences may be part of the reason you are avoiding him) but he is trying to
see you. So he values Party 2 at 10, party 1 at 5 and your presence at the party he
attends is worth 10. These values are reflected in the following table.
Table 7-24: Avoiding Rocky
Rocky
Party 1 Party 2
Party 1 (5,i5) (20,10)
Party 2 (15,5) (0,20)
O
McAfee: Introduction to Economic Analysis, http://www.introecon.com, July 24, 2006
7.264
7.1.4.4 (Exercise) (i) Show there are no pure strategy Nash equilibria in this game.
(ii) Find the mixed strategy Nash equilibria. (iii) Show that the probability you
encounter Rocky is 12
Our final example involves two firms competing for customers. These firms can either
price high or low. The most money is made if they both price high, but if one prices low,
it can take most of the business away from the rival. If they both price low, they make
modest profits. This description is reflected in the following table:
Table 7-25: Price Cutting Game
Firm 2
High Low
S High (15,15) (0,25)
Low (25,0) (5,5)
7.1.4.5 (Exercise) Show that the firms have a dominant strategy to price low, so that
the only Nash equilibrium is (Low, Low).
7.1.5 Two Period Games
So far, we have considered only games that are played simultaneously. Several of these
games, notably the price cutting and apartment cleaning games, are actually played over
and over again. Other games, like the bank location game, may only be played once but
nevertheless are played over time. Recall the bank location game:
Table 7-26; Bank Location Revisited
NYC
No Concession Tax Rebate
No Concession (30,10) (10,20)
Tax Rebate (20,10) (20,0)
If neither city offered a rebate, then LA won the bidding. So suppose instead of the
simultaneous move game, that first New York decided whether to offer a rebate, and
then LA could decide to offer a rebate. This sequential structure leads to a game that
looks like Figure 7-i:
In this game, NYC makes the first move, and chooses Rebate (to the left) or No Rebate
(to the right). If NYC chooses Rebate, LA can then choose Rebate or None. Similarly, if
NYC chooses No Rebate, LA can choose Rebate or None. The payoffs (using the
standard of (LA, NYC) ordering) are written below the choices.
McAfee: Introduction to Economic Analysis, http://www.introecon.com, July 24, 2006
7-265
Rel
Figure 7-1 Sequential Bank Location (NYC payoff listed first)
What NYC would like to do depends on what NYC believes LA will do. What should
NYC believe about LA? (Boy does that rhetorical question suggest a lot of facetious
answers.) The natural belief is that LA will do what is in LA's best interest. This idea -
that each stage of a game is played in a maximizing way - is called subgame perfection.
7.1.6 Subgame Perfection
Subgame perfection requires each player to act in its best interest, independent of the
history of the game.90 This seems very sensible and in most contexts it is sensible. In
some settings, it may be implausible. Even if I see a player make a particular mistake
three times in a row, subgame perfection requires that I must continue to believe that
player will not make the mistake again. Subgame perfection may be implausible in
some circumstances, especially when it pays to be considered somewhat crazy.
In the example, subgame perfection requires LA to offer a rebate when NYC does (since
LA gets 20 by rebating versus 10), and not when NYC doesn't. This is illustrated in the
game using arrows to indicate LA's choices. In addition, the actions that LA won't
choose have been re-colored in a light grey in Figure 7-2.
Once LA's subgame perfect choices are taken into account, NYC is presented with the
choice of offering a rebate, in which case it gets o, or not offering a rebate, in which case
it gets 10. Clearly the optimal choice for NYC is to offer no rebate, in which case LA
doesn't either, and the result is 30 for LA, and 10 for NYC.
Dynamic games are generally "solved backward" in this way. That is, first establish what
the last player does, then figure out based on the last player's expected behavior, what
the penultimate player does, and so on.
90 Subgame perfection was introduced by Nobel laureate Reinhart Selten (1930 - ).
McAfee: Introduction to Economic Analysis, http://www.introecon.com, July 24, 2006
7.266
No Rebate
'LA
Rebat
(20,0)
(30,10)
Figure 7-2: Subgame Perfection
We'll consider one more application of subgame perfection. Suppose, in the game
"Avoiding Rocky," Rocky is actually stalking you, and can condition his choice on your
choice. Then you might as well go to the party you like best, because Rocky is going to
follow you wherever you go. This is represented in Figure 7-3.
ou
Party 1 Party2
R R
Party 1 arty 2
(5,15) (0,20)
Figure 7-3: Can't Avoid Rocky
Since Rocky's optimal choice eliminates your best outcomes, you make the best of a bad
situation by choosing Party 1. Here, Rocky has a second mover advantage: Rocky's
ability to condition on your choice meant he does better than he would do in a
simultaneous game.
7.1.6.1 (Exercise) Formulate the battle of the sexes as a sequential game, letting the
woman choose first. (This situation could arise if the woman can leave a
message for the man about where she has gone.) Show that there is only one
McAfee: Introduction to Economic Analysis, http://www.introecon.com, July 24, 2006
7-267
subgame perfect equilibrium, and that the woman enjoys a first-mover
advantage over the man, and she gets her most preferred outcome.
7.1.7 Supergames
Some situations, like the pricing game or the apartment-cleaning game, are played over
and over. Such situations are best modeled as a supergame.91 A supergame is a game
played over and over again without end, where the players discount the future. The
game played each time is known as a stage game. Generally supergames are played in
times 1, 2, 3, ...
Cooperation may be possible in supergames, if the future is important enough.
Consider the pricing game introduced above.
Table 7-27: Price Cutting, Revisited
Firm 2
High Low
S High (15,15) (0,25)
Low (25,0) (5,5)
The dominant strategy equilibrium to this game is (Low, Low). It is clearly a subgame
perfect equilibrium for the players to just play (Low, Low) over and over again, because
if that is what Firm 1 thinks Firm 2 is doing, Firm 1 does best by pricing Low, and vice
versa. But that is not the only equilibrium to the supergame.
Consider the following strategy, called a grim trigger strategy. Price high, until you see
your rival price low. After your rival has priced low, price low forever. This is called a
trigger strategy because an action of the other player (pricing low) triggers a change in
behavior. It is a grim strategy because it punishes forever.
If your rival uses a grim trigger strategy, what should you do? Basically, your only
choice is when to price low, because once you price low, your rival will price low, and
then your best choice is to also price low from then on. Thus, your strategy is to price
high up until some point t - 1, and then price low from time t on. Your rival will price
high through t, and price low from t + 1 on. This gives a payoff to you of 15 from period
1 through t - 1, 25 in period t, and then 5 in period t + 1 on. We can compute the payoff
for a discount factor 6.
V, = 15(1+6+ ..+6t-1)+256' + 5(6'* +2..
-15 + 25V'+5 =~ 15 (15 -_5(-__-5 - (-10 +206).
1-6 1-6 1-6 1-6 1-6 1-6
91 The supergame was invented by Robert Aumann (1930 - ) in 1959.
McAfee: Introduction to Economic Analysis, http://www.introecon.com, July 24, 2006
7.268
If -10 + 206 < o, it pays to price low immediately, at t=o, because it pays to price low
and the earlier the higher the present value. If -10 + 206 > o, it pays to wait forever to
price low, that is, t = co. Thus, in particular, the grim trigger strategy is an optimal
strategy for a player when the rival is playing the grim trigger strategy if 6 1/2. In other
words, cooperation in pricing is a subgame perfect equilibrium if the future is important
enough, that is, the discount factor 6 is high enough.
The logic of this example is that the promise of future cooperation is valuable when the
future itself is valuable, and that promise of future cooperation can be used to induce
cooperation today. Thus, firm 1 doesn't want to cut price today, because that would lead
firm 2 to cut price for the indefinite future. The grim trigger strategy punishes price
cutting today with future low profits.
Supergames offer more scope for cooperation than is illustrated in the pricing game.
First, more complex behavior is possible. For example, consider the following game:
Table 7-28: A Variation of the Price Cutting Game
Firm 2
High Low
High (10,10) (0,25)
Low (25,0) (5,5)
Here, again, the unique equilibrium in the stage game is (Low, Low). But the difference
between this game and the previous game is that the total profits of firms 1 and 2 are
higher in either (High, Low) or (Low, High) than in (High, High). One solution is to
alternate between (High, Low) and (Low, High). Such alternation can also be supported
as an equilibrium, using the grim trigger strategy - that is, if a firm does anything other
than what is it supposed to in the alternating solution, the firms instead play (Low, Low)
forever.
7.1.7.1 (Exercise) Consider the game in Table 7-28, and consider a strategy in which
firm 1 prices high in odd numbered periods, and low in even numbered periods,
while 2 prices high in even numbered periods, low in odd numbered periods. If
either deviate from these strategies, both price low from then on. Let 6 be the
discount factor. Show that these firms have a payoff of 25 or 256
1-62 1-_2
depending on which period it is. Then show that the alternating strategy is
sustainable if 10+ 56 256 This, in turn, is equivalent to 6 2J -2.
1-6 1_62
7.1.8 The Folk Theorem
The folk theorem says that if the value of the future is high enough, any outcome that is
individually rational can be supported as an equilibrium to the supergame. Individual
rationality for a player in this context means that the outcome offers a present value of
profits at least as high as that offered in the worst equilibrium in the stage game from
that player's perspective. Thus, in the pricing game, the worst equilibrium of the stage
McAfee: Introduction to Economic Analysis, http://www.introecon.com, July 24, 2006
7.269
game offered each player 5, so an outcome can be supported if it offers each player at
least a running average of 5.
The simple logic of the folk theorem is this. First, any infinite repetition of an
equilibrium of the stage game is itself a subgame perfect equilibrium. If everyone
expects this repetition of the stage game equilibrium, no one can do better than to play
their role in the stage game equilibrium every period. Second, any other plan of action
can be turned into a subgame perfect equilibrium merely by threatening any agent who
deviates from that plan with an infinite repetition of the worst stage game equilibrium
from that agent's perspective. That threat is credible because the repetition of the stage
game equilibrium is itself a subgame perfect equilibrium. Given such a grim trigger type
threat, no one wants to deviate from the intended plan.
The folk theorem is a powerful result, and shows that there are equilibria to supergames
that achieve very good outcomes. The kinds of coordination failures we saw in the battle
of the sexes, and the failure to cooperate in the prisoner's dilemma, need not arise, and
cooperative solutions are possible if the future is sufficiently valuable.
However, it is worth noting some assumptions that have been made in our descriptions
of these games, assumptions that matter and are unlikely to be true in practice. First,
the players know their own payoffs. Second, they know their rival's payoffs. They
possess a complete description of the available strategies and can calculate the
consequences of these strategies, not just for themselves, but for their rivals. Third,
each player maximizes his or her expected payoff, and they know that their rivals do the
same, and they know that their rivals know that everyone maximizes, and so on. The
economic language for this is the structure of the game and the player's preferences are
common knowledge. Few real world games will satisfy these assumptions exactly.
Since the success of the grim trigger strategy (and other strategies we haven't discussed)
generally depends on such knowledge, informational considerations may cause
cooperation to break down. Finally, the folk theorem shows us that there are lots of
equilibria to supergames, and provides no guidance on which one will be played. These
assumptions can be relaxed, although they may lead to wars on the equilibrium path "by
accident," and a need to recover from such wars, so that the grim trigger strategy
becomes sub-optimal.
7.2 Cournot Oligopoly
The Cournot92 oligopoly model is the most popular model of imperfect competition. It is
a model in which the number of firms matters, and represents one way of thinking about
what happens when the world is neither perfectly competitive, nor a monopoly.
In the Cournot model, there are n firms, who choose quantities. We denote a typical
firm as firm i and number the firms from i = 1 to i = n. Firm i chooses a quantity qi 0
to sell and this quantity costs cj(qi). The sum of the quantities produced is denoted by Q.
The price that emerges from the competition among the firms is p(Q) and this is the
same price for each firm. It is probably best to think of the quantity as really
92 Augustus Cournot, 1801-1877.
McAfee: Introduction to Economic Analysis, http://www.introecon.com, July 24, 2006
7.270
representing a capacity, and competition in prices by the firms determining a market
price given the market capacity.
The profit that a firm i obtains is
nri = p(Q)qi - ci(qi ).
7.2.1 Equilibrium
Each firm chooses qi to maximize profit. The first order conditions93 give:
o - 1 -p(Q)+ p'(Qqi - c (qi)
aqi
This equation holds with equality provided qi > o. A simple thing that can be done with
the first order conditions is to rewrite them to obtain the average value of the price-cost
margin:
p(Q)-c (qji) _p'(Q)qi Qp'(Q) qi si
p(Q) p(Q) p(Q) Q *
Here si = - is firm i's market share. Multiplying this equation by the market share and
Q
summing over all firms i = 1, ... , n yields
" p(Q) - ci(qi) 1 2 HHI
_ p(Q) Ei_1 E
n
where HHI = $s is the Hirschman-Herfindahl Index.94 The HHI has the property
i=1
that if the firms are identical, so that si = 1/n for all i, then the HHI is also 1/n. For this
reason, antitrust economists will sometimes use 1/HHI as a proxy for the number of
firms, and describe an industry with "2 1/2 firms," meaning an HHI of 0.4.95
We can draw several inferences from these equations. First, larger firms, those with
larger market shares, have a larger deviation from competitive behavior (price equal to
marginal cost). Small firms are approximately competitive (price nearly equals
marginal cost) while large firms reduce output to keep the price higher, and the amount
of the reduction, in price/cost terms, is proportional to market share. Second, the H HI
reflects the deviation from perfect competition on average, that is, it gives the average
93 Bear in mind that Q is the sum of the firms' quantities, so that when firm i increases its output slightly,
Q goes up by the same amount.
94 Named for Albert Hirschman (1918 - 1972), who invented it in 1945, and Orris Herfindahi (1915 - ),
who invented it independently in 1950.
95 To make matters more confusing, antitrust economists tend to state the H HI using shares in percent, so
that the H HI is on a 0 to 10,000 scale.
McAfee: Introduction to Economic Analysis, http://www.introecon.com, July 24, 2006
7.271
proportion by which price equal to marginal cost is violated. Third, the equation
generalizes the "inverse elasticity result" proved for monopoly, which showed that the
price - cost margin was the inverse of the elasticity of demand. The generalization
states that the weighted average of the price - cost margins is the HHI over the elasticity
of demand.
Since the price - cost margin reflects the deviation from competition, the HHI provides
a measure of how large a deviation from competition is present in an industry. A large
HHI means the industry "looks like monopoly." In contrast, a small HHI looks like
perfect competition, holding constant the elasticity of demand.
The case of a symmetric (identical cost functions) industry is especially enlightening. In
this case, the equation for the first order condition can be restated as
o = p(Q) +p'(Q)Q - c' Q
or
p(Q)= En c' .
En-1 n
Thus, in the symmetric model, competition leads to pricing as if demand were more
elastic, and indeed is a substitute for elasticity as a determinant of price.
7.2.2 Industry Performance
How does the Cournot industry perform? Let us return to the more general model, that
doesn't require identical cost functions. We already have one answer to this question:
the average price - cost margin is the HHI divided by the elasticity of demand. Thus, if
we have an estimate of the demand elasticity, we know how much the price deviates
from the perfect competition benchmark.
The general Cournot industry actually has two sources of inefficiency. First, price is
above marginal cost, so there is the dead weight loss associated with unexploited gains
from trade. Second, there is the inefficiency associated with different marginal costs.
This is inefficient because a re-arrangement of production, keeping total output the
same, from the firm with high marginal cost to the firm with low marginal cost, would
reduce the cost of production. That is, not only is too little output produced, but what
output is produced is inefficiently produced, unless the firms are identical.
To assess the productive inefficiency, we let c' be the lowest marginal cost. The average
deviation from the lowest marginal cost, then, is
n n n
S= Esi(c) -c') = Esi(p -c' -(p -ci)) = p -c' - Esi(p -ci)
i=1 i=1 i=1
McAfee: Introduction to Economic Analysis, http://www.introecon.com, July 24, 2006
7.272
" (p-cf)
=p-c'-pjsi p = p - s=p-c' HHI.
i=1 P si=1
Thus, while a large HHI means a large deviation from price equal to marginal cost and
hence a large level of monopoly power (holding constant the elasticity of demand), a
large HHI also tends to indicate greater productive efficiency, that is, less output
produced by high cost producers. Intuitively, a monopoly produces efficiently, even if it
has a greater reduction in total output than other industry structures.
There are a number of caveats worth mentioning in the assessment of industry
performance. First, the analysis has held constant the elasticity of demand, which could
easily fail to be correct in an application. Second, fixed costs have not been considered.
An industry with large economies of scale, relative to demand, must have very few firms
to perform efficiently and small numbers should not necessarily indicate the market
performs poorly even if price - cost margins are high. Third, it could be that entry
determines the number of firms, and that the firms have no long-run market power, just
short-run market power. Thus, entry and fixed costs could lead the firms to have
approximately zero profits, in spite of price above marginal cost.
7.2.2.1 (Exercise) Suppose the inverse demand curve is p(Q) = 1 - Q, and that there
are n Cournot firms, each with constant marginal cost c, selling in the market.
(i) Show that the Cournot equilibrium quantity and price are Q = n(1- c) and
n+1
1+nc.1-2
p(Q)= . (ii) Show each firm's gross profits are .
n+1 n+1
Continuing with 7.2.2.1 (Exercise), suppose there is a fixed cost Fto become a firm. The
number of firms n should be such that firms are able to cover their fixed costs, but add
one more and they can't. This gives us a condition determining the number of firms n:
1- 2F
n+1)j n+2)
Thus, each firm's net profits are 1-C -F (1-C 2 j1-c 2 _ (2n+3)(1-c)2
(n+1) (n+1) (n+2 ) (n+1)2(n+2)2
Note that the monopoly profits arm are 1/ (1-c)2. Thus, with free entry, net profits are
(2n +3)4 .n(2n +3)4
less than xcm, and industry net profits are less than a
(n +1)2 (n +2)2 (n +1)2(n +2)2
Table 7-29 shows the performance of the constant cost, linear demand Cournot
industry, when fixed costs are taken into account, and when they aren't. With two firms,
gross industry profits are 8/9ths of the monopoly profits, not substantially different
from monopoly. But when fixed costs sufficient to insure that only two firms enter are
McAfee: Introduction to Economic Analysis, http://www.introecon.com, July 24, 2006
7.273
considered, the industry profits are at most 39% of the monopoly profits. This number
- 39% -- is large because fixed costs could be "relatively" low, so that the third firm is
just deterred from entering. That still leaves the two firms with significant profits, even
though the third firm can't profitably enter. As the number of firms rises, gross industry
profits fall slowly toward zero. The net industry profits, on the other hand, fall
dramatically rapidly to zero. With ten firms, the gross profits are still about a third of
the monopoly level, but the net profits are only at most 5% of the monopoly level.
Table 7-29: Industry Profits as a Fraction of Monopoly Profits
Number Gross Net
of Firms Industry Industry
Profits (%) Profits (%)
2 88.9 39.0
3 75.0 27.0
4 64.0 19.6
5 55.6 14.7
10 33.1 5.3
15 23.4 2.7
20 18.1 1.6
The Cournot model gives a useful model of imperfect competition, a model that readily
permits assessing the deviation from perfect competition. The Cournot model embodies
two kinds of inefficiency: the exercise of monopoly power, and technical inefficiency in
production. In settings involving entry and fixed costs, care must be taken in applying
the Cournot model.
7.3 Search and Price Dispersion
Decades ago, economists used to make a big deal about the "Law of One Price," which
states that identical goods sell at the same price. The argument in favor of the law of
one price is theoretical. Well-informed consumers will buy identical goods from the
lowest price seller. Consequently, the only seller to make any sales is the low-price
seller. This kind of consumer behavior forces all sellers to sell at the same price.
There are few markets where the law of one price is actually observed to hold.
Organized exchanges, like stock, bond and commodity markets, will satisfy the law of
one price. In addition, gas stations across the street from each other will often offer
identical prices, but often is not always.
Many economists considered that the internet would force prices of standardized goods
- DVD players, digital cameras, MP3 players - to a uniform, and uniformly low, price.
However, this has not occurred. Moreover, it probably can't occur, in the sense that
pure price competition would put the firms out of business, and hence can't represent
equilibrium behavior.
There are many markets where prices appear unpredictable to consumers. The price of
airline tickets is notorious for unpredictability. The price of milk, soft drinks, paper
McAfee: Introduction to Economic Analysis, http://www.introecon.com, July 24, 2006
7.274
towels and canned tuna varies 50% or more depending on whether the store has an
advertised sale of the item or not. Prices of goods sold on the internet varies
substantially from day to day.96 Such variation is known as price dispersion by
economists. It is different from price discrimination, in that price dispersion entails a
given store quoting the same price to all customers; the variation is across stores, while
price discrimination is across customers.
Why are prices so unpredictable?
7.3.1 Simplest Theory
To understand price dispersion, we divide consumers into two types: shoppers and loyal
customers. Loyal customers won't pay more than a price pm for the good, but they only
consult a particular store; if that store has the good for less than the price pm, the loyal
customer buys, and otherwise not. In contrast, the shoppers buy only from the store
offering the lowest price; shoppers are informed about the prices offered by all stores.
We let the proportion of shoppers be s. The loyal customers are allocated to the other
stores equally, so that, if there are n stores, each store gets a fraction (1 - s)/n of the
customers. Let the marginal cost of the good be c, and assume c < pm. Both kinds of
customers buy only one unit.
For the purposes of this analysis, we will assume that prices can be chosen from the
continuum. This makes the analysis more straightforward, but there is an alternate
version of the analysis (not developed here) that makes the more reasonable assumption
of prices that are an integer number of pennies.
First note that there is no pure strategy equilibrium. To see this, consider the lowest
price p charged by any firm. If that price is c, the firm makes no money, so would do
better by raising its price to pm and selling only to the loyal customers. Thus, the lowest
price p exceeds c. If there is a tie at p, it pays to break the tie by charging a billionth of a
cent less than p, and thereby capturing all the shoppers rather than sharing them with
the other firm charging p. So there can't be a tie.
But no tie at p means the next lowest firm is charging something strictly greater than p,
which means the lowest price firm can increase price somewhat and not suffer any loss
of sales. This contradicts profit maximization for that firm. The conclusion is that firms
must randomize and no pure strategy equilibrium exists.
But how do they randomize? We are going to look for a distribution of prices. Each firm
will choose a price from the continuous distribution F, where F(x) is the probability the
firm charges a price less than x. What must F look like? We use the logic of mixed
strategies: the firm must get the same profits for all prices that might actually be
charged under the mixed strategy, for otherwise it would not be willing to randomize.
A firm that charges price p Pm always sells to its captive customers. In addition, it sells
to the shoppers if the other firms have higher prices, which occurs with probability
(1 - F(p))"-1. Thus, the firm's profits are
96 It is often very challenging to assess internet prices because of variation in shipping charges.
McAfee: Introduction to Economic Analysis, http://www.introecon.com, July 24, 2006
7-275
1(p)= (p - c) 1 s + s(i- F(p))"1.
n)
On each sale, the firm earns p - c. The firm always sells to its loyal customers, and in
addition captures the shoppers if the other firms price higher. Since no firm will exceed
Pm, the profits must be the same as the level arising from charging pm, and this gives
(p)=(P -c)-1s s(1-sF(p))" =(Pm -c)1ns
This equation is readily solved for F:
F(p) = 1 (pm p)(-s)n
s(p-c)n ,
The lower bound of prices arises at the point L where F(L)=o, or
(pm -c)1-s
L=c+ n
1-s
+s
n
These two equations provide a continuous distribution of prices charged by each firm
which is an equilibrium to the pricing game. That is, each firm randomizes over the
interval [L, pm], according to the continuous distribution F. Any price in the interval
[L,pm] produces the same profits for each firm, so the firms are willing to randomize
over this interval.
The loyal customers get a price chosen randomly from F, so we immediately see that the
shoppers make life better for the loyal customers, pushing average price down. (An
increase in s directly increases F, which means prices fall - recall that F gives the
probability that prices are below a given level, so an increase in F is an increase in the
probability of low prices.)
Similarly loyal customers make life worse for shoppers, increasing prices on average to
shoppers. The distribution of prices facing shoppers is actually the distribution of the
minimum price. Since all firms charge a price exceeding p with probability (1 -Fp),
at least one charges a price less than p with probability 1 - (1 - F(p))n, and this is the
distribution of prices facing shoppers. That is, the distribution of prices charged to
shoppers is
1-(- ~p)"=1-K (Pm -p)(1s) -
McAfee: Introduction to Economic Analysis, http://www.introecon.com, July 24, 2006
7.276
7.3.2 Industry Performance
How does a price dispersed industry perform? First, average industry profits are
nc(p) = (pm - c)(1- s).
An interesting aspect of this equation is that it doesn't depend on the number of firms,
only on the number of loyal customers. Essentially, the industry profits are the same
that it would earn as if the shoppers paid marginal cost and the loyal customers paid the
monopoly price, although that isn't what happens in the industry, except in the limit as
the number of firms goes to infinity. Note that this formula for industry profits does not
work for a monopoly. In order to capture monopoly, one must set s=o, because
shoppers have no alternative under monopoly.
Price
1r7
0.8
0.6
0.4
0.2
Expected price for loyal
customers
Expected price for shoppers
n
20 40
60 80 100
Figure 7-4: Expected Prices in Search Equilibrium
As the number of firms gets large, the price charged by any one firm converges to the
monopoly price pm. However, the lowest price offered by any firm actually converges to
c, marginal cost. Thus, in the limit as the number of firms get large, shoppers obtain
price equal to marginal cost and loyal firms pay the monopoly price.
The average price charged to shoppers and non-shoppers is a complicated object, so we
consider the case where there are n firms, s = 1/2, pm =1 and c = o. Then the expected
prices for shoppers and loyal customers are given in Figure 7-4, letting the number of
firms vary. Thus, with many firms, most of the gains created by the shoppers flow to
shoppers. In contrast, with few firms, a significant fraction of the gains created by
shoppers goes instead to the loyal customers.
McAfee: Introduction to Economic Analysis, http://www.introecon.com, July 24, 2006
7.277
Similarly, we can examine the average prices for loyal customers and shoppers when the
proportion of shoppers varies. Increasing the proportion of shoppers has two effects.
First, it makes low prices more attractive, thus encouraging price competition, because
capturing the shoppers is more valuable. Second, it lowers industry profits, because the
set of loyal customers is reduced. Figure 7-5 plots the average price for loyal customers
and shoppers, as the proportion of shoppers ranges from zero to one, when there are
five firms, pm = 1 and c = o.
Price
1
Expected price for
loyal customers
0.8
0.6
0.4
0.2
Expected price
for shoppers
S
0.2 0.4 0.6 0.8 1
Figure 7-5: Expected Prices (s=Proportion of Shoppers)
People who are price-sensitive and shop around convey a positive externality on other
buyers by encouraging price competition. Similarly, people who are less price sensitive
and don't shop around convey a negative externality on the other buyers. In markets
with dispersed information about the best prices, where some buyers are informed and
some are not, randomized prices are a natural outcome. That is, randomization of
prices, and the failure of the law of one price, is just a reflection of the different
willingness or ability to search on the part of consumers.
This difference in the willingness to search could arise simply because search is itself
costly. That is, the shoppers could be determined by their choice to shop, in such a way
that the cost of shopping just balances the expected gains from searching. The
proportion of shoppers may adjust endogenously to insure that the gains from searching
exactly equal the costs of searching. In this way, a cost of shopping is translated into a
randomized price equilibrium in which there is a benefit from shopping and all
consumers get the same total cost of purchase on average.
McAfee: Introduction to Economic Analysis, http://www.introecon.com, July 24, 2006
7.278
7.4 Hotelling Model
Breakfast cereals range from indigestible, unprocessed whole grains to cereals that are
almost entirely sugar with only the odd molecule or two of grain. Such cereals are
hardly good substitutes for each other. Yet similar cereals are viewed by consumers as
good substitutes, and the standard model of this kind of situation is the Hotelling
model.97 Hotelling was the first to use a line segment to represent both the product that
is sold and the preferences of the consumers who are buying the products. In the
Hotelling model, there is a line, and preferences of each consumer is represented by a
point on this line. The same line is used to represent products. For example, movie
customers are differentiated by age, and we can represent moviegoers by their ages.
Movies, too, are designed to be enjoyed by particular ages. Thus a "pre-teen" movie is
unlikely to appeal very much to a six year old or to a nineteen year old, while a Disney
movie appeals to a six year old, but less to a fifteen year old. That is, movies have a
target age, and customers have ages, and these are graphed on the same line.
High Fiber Adult Cereals Kid Cereals
Sugar Content
Figure 7-6: Hotelling Model for Breakfast Cereals
Breakfast cereal is a classic application of the Hotelling line, and this application is
illustrated in Figure 7-6. Breakfast cereals are primarily distinguished by their sugar
content, which ranges on the Hotelling line from low on the left to high on the right.
Similarly, the preferences of consumers also fall on the same line. Each consumer has a
"most desired point," and prefers cereals closer to that point than those at more distant
points.
7.4.1 Types of Differentiation
There are two main types of differentiation, each of which can be modeled using the
Hotelling line. These types are quality and variety. Quality refers to a situation where
consumers agree what product is better; the disagreement among consumers concerns
whether higher quality is worth the cost. In automobiles, faster acceleration, better
braking, higher gas mileage, more cargo capacity, more legroom, and greater durability
are all good things. In computers, faster processing, brighter screens, higher resolution
screens, lower heat, greater durability, more megabytes of RAM and more gigabytes of
hard drive space are all good things. In contrast, varieties are the elements about which
there is not widespread agreement. Colors and shapes are usually varietal rather than
quality differentiators. Some people like almond colored appliances, others choose
white, with blue a distant third. Food flavors are varieties, and while the quality of
ingredients is a quality differentiator, the type of food is usually a varietal differentiator.
Differences in music would primarily be varietal.
Quality is often called vertical differentiation, while variety is horizontal
differentiation.
97 Hotelling Theory is named for Harold Hotelling, 1895-1973.
McAfee: Introduction to Economic Analysis, http://www.introecon.com, July 24, 2006
7.279
7.4.2 The Standard Model
The standard Hotelling model fits two ice cream vendors on a beach. The vendors sell
the identical product, and moreover they can choose to locate wherever they wish. For
the time being, suppose the price they charge for ice cream is fixed at $1. Potential
customers are also spread randomly along the beach.
We let the beach span an interval from o to 1. People desiring ice cream will walk to the
closest vendor, since the price is the same. Thus, if one vendor locates at x and the other
at y, and x < y, those located between o and 1/2 (x + y) go to the left vendor, while the
rest go to the right vendor. This is illustrated in Figure 7-7.
Buy atx Buy aty
o %2(+y)
Figure 7-7: Sharing the Hotelling Market
Note that the vendor at x sells more by moving toward y, and vice versa. Such logic
forces profit maximizing vendors to both locate in the middle! The one on the left sells
to everyone left of 1/2, while the one on the right sells to the rest. Neither can capture
more of the market, so equilibrium locations have been found. (To complete the
description of an equilibrium, we need to let the two "share" a point and still have one
on the right side, one on the left side of that point.)
This solution is commonly used as an explanation of why U.S. political parties often
seem very similar to each other - they have met in the middle in the process of chasing
the most voters. Political parties can't directly buy votes, so the "price" is fixed; the only
thing parties can do is locate their platform close to voters' preferred platform, on a
scale of "left" to "right." But the same logic that a party can grab the middle, without
losing the ends, by moving closer to the other party will tend to force the parties to share
the same "middle of the road" platform.
7.4.2.1 (Exercise) Suppose there are four ice cream vendors on the beach, and
customers are distributed uniformly. Show that it is a Nash equilibrium for two
to locate at 1/4, and two at 3/4.
The model with constant prices is unrealistic for the study of the behavior of firms.
Moreover, the two-firm model on the beach is complicated to solve and has the
undesirable property that it matters significantly whether the number of firms is odd or
even. As a result, we will consider a Hotelling model on a circle, and let the firms choose
their prices.
7.4.3 The Circle Model
In this model, there are n firms evenly spaced around the circle whose circumference is
one. Thus, the distance between any firm and each of its closest neighbors is 1/n.
McAfee: Introduction to Economic Analysis, http://www.introecon.com, July 24, 2006
7.280
Consumers care about two things: how distant the firm they buy from is, and how much
they pay for the good, and they minimize the sum of the price paid and t times the
distance between the consumer's location (also on the circle) and the firm. Each
consumer's preference is uniformly distributed around the circle. The locations of firms
are illustrated in Figure 7-8.
length 1/n length 1/n
Firm Firm Firm
Figure 7-8: A Segment of the Circle Model
We conjecture a Nash equilibrium in which all firms charge the price p. To identify p,
we look for what p must be to make any one firm choose to charge p, given that the
others all charge p. So suppose the firm in the middle of Figure 7-8 charges an alternate
price r, but every other firm charges p. A consumer who is x units away from the firm
pays the price r + tx from buying at the firm, or p + t(1/n - x) from buying from the
rival. The consumer is just indifferent between the nearby firms if these are equal, that
is,
r + tx* = p + t(1/n - x*)
where x* is the location of the consumer who is indifferent.
, p+n-r 1+p-r
2t 2n 2t
Thus, consumers who are closer than x* to the firm charging r buy from that firm, and
consumers who are further away than x* buy from the alternative firm. Demand for the
firm charging r is twice x* (because the firm sells to both sides), so profits are price
minus marginal cost times two x*, that is,
(r - c)2x* = (r - c) -+ p-r
n t
The first order condition98 for profit maximization is
a 1_p-r 1p-r r-c
0 = (r -c) -+P> -+Pj- t
98 SinCe profit is quadratiC in r, we will find a global maXimUm.
McAfee: Introduction to Economic Analysis, http://www.introecon.com, July 24, 2006
7.281
We could solve the first order condition for r. But remember that the concern is when is
p a Nash equilibrium price? The price p is an equilibrium price if the firm wants to
choose r = p. Thus, we can conclude that p is a Nash equilibrium price when
t
p=c+-.
n
This value of p insures that a firm facing rivals who charge p also chooses to charge p.
Thus, in the Hotelling model, price exceeds marginal cost by an amount equal to the
value of the average distance between the firms, since the average distance is 1/n and
the value to a consumer of traveling that distance is t. The profit level of each firm is
t t
, so industry profits are -.
n2 n
How many firms will enter the market? Suppose the fixed cost is F. We are going to
take a slightly unusual approach and assume that the number of firms can adjust in a
continuous fashion, in which case the number of firms is determined by the zero profit
condition F = ,or n= .
n2
What is the socially efficient number of firms? The socially efficient number of firms
minimizes the total costs, which are the sum of the transportation costs and the fixed
costs. With n firms, the average distance a consumer travels is
2n xdx=2n Jxdx=n J= .
2) 4n
-2n 0
Thus, the socially efficient number of firms minimizes the transport costs plus the entry
costs +nF. This occurs at n = . The socially efficient number of firms is half
the level that enter with free entry!
Too many firms enter in the Hotelling circle model. This extra entry arises because
efficient entry is determined by the cost of entry and the average distance of consumers,
while prices are determined by the marginal distance of consumers, or the distance of
the marginal consumer. That is, competing firms' prices are determined by the most
distant customer, and that leads to prices that are too high relative to the efficient level;
free entry then drives net profits to zero only by excess entry.
The Hotelling model is sometimes used to justify an assertion that firms will advertise
too much, or engage in too much R&D, as a means of differentiating themselves and
creating profits.
McAfee: Introduction to Economic Analysis, http://www.introecon.com, July 24, 2006
7.282
7.5 Agency Theory
An agent is a person who works for, or on behalf of, another. Thus, an employee is an
agent of a company. But agency extends beyond employee relationships. Independent
contractors are also agents. Advertising firms, lawyers and accountants are agents of
their clients. The CEO of a company is an agent of the board of directors of the
company. A grocery store is an agent of the manufacturer of corn chips sold in the store.
Thus, the agency relationship extends beyond the employee into many different
economic relationships. The entity - person or corporation - on whose behalf an agent
works is called a principal.
Agency theory is the study of incentives provided to agents. Incentives are an issue
because agents need not have the same interests and goals as the principal. Employees
spend billions of hours every year browsing the web, emailing friends, and playing
computer games while they are supposedly working. Attorneys hired to defend a
corporation in a lawsuit have an incentive not to settle, to keep the billing flowing.
(Such behavior would violate the attorneys' ethics requirements.) Automobile repair
shops have been known to use cheap or used replacement parts and bill for new, high
quality parts. These are all examples of a conflict in the incentives of the agent and the
goals of the principal.
Agency theory focuses on the cost of providing incentives. When you rent a car, an
agency relationship is created. Even though a car rental company is called an agency, it
is most useful to look at the renter as the agent, because it is the renter's behavior that is
an issue. The company would like the agent to treat the car as if it were their own car.
The renter, in contrast, knows it isn't their own car, and often drives accordingly.
"[T]here's a lot of debate on this subject---about what kind of car handles
best. Some say a front-engined car; some say a rear-engined car. I say a
rented car. Nothing handles better than a rented car. You can go faster, turn
corners sharper, and put the transmission into reverse while going forward at
a higher rate of speed in a rented car than in any other kind." 99
How can the car rental company insure that you don't put their car into reverse while
going forward at a high rate of speed? They could monitor your behavior, perhaps by
putting a company representative in the car with you. That would be a very expensive
and unpleasant solution to the problem of incentives. Instead, the company uses
outcomes - if damage is done, the driver has to pay for it. That is also an imperfect
solution, because some drivers who abuse the cars get off scot-free and others who don't
abuse the car still have cars that break down, and are then mired in paperwork while
they try to prove their good behavior. That is, a rule that penalizes drivers based on
outcomes imposes risk on the drivers. Modern technology is improving monitoring with
GPS tracking.
99 P. J. O'Rourke, Republican Party Reptile, Atlantic Monthly Press, 1987.
McAfee: Introduction to Economic Analysis, http://www.introecon.com, July 24, 2006
7.283
7.5.1 Simple Model
To model the cost of provision of incentives, we consider an agent like a door-to-door
encyclopedia salesperson. The agent will visit houses, and sell encyclopedias to some
proportion of the households; the more work the agent does, the more sales that are
made. We let x represent the average dollar value of sales for a given level of effort; x is
a choice the agent makes. However, x will come with risk to the agent, which we model
using the variance 62.
The firm will pay the agent a share s of the money generated by the sales. In addition,
the firm will pay the agent a salary y, which is fixed independently of sales. This scheme
- a combination of salary and commission - covers many different situations. Real
estate agents receive a mix of salary and commission. Authors receive an advance and a
royalty, which works like a salary and commission.
The monetary compensation of the agent is sx + y. In addition, the agent has a cost of
x2
effort, which we take to be . Here a represents the ability of the agent: more able
2a
agents, who have a higher value of a, have a lower cost of effort. Finally, there is a cost
of risk. The actual risk imposed on the agent is proportional to the degree they share in
the proceeds; if s is small, the agent faces almost no monetary risk, while if s is high,
most of the risk is imposed on the agent. We use the "linear cost of risk" model
developed earlier, to impose a cost of risk which is s),a2. Here, 62 is the variance of the
monetary risk, 2, defines the agent's attitude or cost of risk, and s is the share of the risk
imposed on the agent. This results in a payoff to the agent of
X2
u=sx+y- -sca2.
2a
The first two terms, sx + y, are the payments made to the agent. The next term is the
cost of generating that level of x. The final term is the cost of risk imposed on the agent
by the contract.
The agency game works as follows. First, the principal offers a contract, which involves
a commission s and a salary y. The agent can either accept or reject the contract and
accepts if he obtains at least uo units of utility, the value of his next best offer. Then the
agent decides how much effort to expend, that is, the agent chooses x.
As with all subgame perfect equilibria, we work backwards to first figure out what x an
agent would choose. Because our assumptions make u quadratic in x, this is a
straightforward exercise, and we conclude x=sa. This can be embedded into u, and we
obtain the agent's optimized utility, u*,i
u*=s2a~y -(sa)2 -s2a2 = y±+/s2a -s2
2a
McAfee: Introduction to Economic Analysis, http://www.introecon.com, July 24, 2006
7.284
Incentive Compensation: A Percentage of What?
Most companies compensate their sales force based on the revenue generated. However, maximizing
revenue need not be the same thing as maximizing profit, which is generally the goal of the company. In
what follows, Steve Bisset discusses the difference.
"Many years ago I was CEO of a company called Megatest, founded by Howard Marshall and myself in
1975. Around 1987 we were selling test systems in the $1M+ price range. Every Monday morning we would
have a meeting with sales management and product marketing, mediated by myself. Individual salesmen
came in to make their cases for how they just had to have a huge discount, else they would lose a particular
sale. The meeting was part circus, with great performances, and part dogfight.
"I could visualize the sales guys spending their time in the car or the shower using their substantial creative
powers to dream up good justifications for their next plea for a huge discount. They knew that if we were
willing to bleed enough we could usually win the sale. I wanted to solve both the resultant profitability
problem and the problem of the unpleasant meeting.
"Commissions were traditionally a percentage of bookings (net of discount), with part held back until cash
was received. The percentage increased after a salesman met his individual quota for the quarter (the
performances at quota-setting time to sandbag the quota were even more impressive). The fact that a
discount reduced commission did not affect a salesman's behavior, because the difference in commission
was small. Better to buy the order by giving a big discount and then move on quickly to the next sale.
"Salesmen are "coin operated", and will figure out how to maximize their total commission. Most salesmen
I have met are quite good at math. Further, they have learned to "watch the hips, not the lips" - in other
words, they judge management intentions by actions rather than words. They reason - and I agree with
them - that if management really wanted them to maximize margins rather than revenues, then they would
pay them more if they maximize margins.
"We decided to try a new scheme, against some howling from the sales department. We set a base "cost"
for each product, approximately representing the incremental cost to manufacture and support the product.
Then we offered commission on the amount that the net sales price exceeded this base cost. The base cost
may have been as much as 80% of the list price (it was a very competitive market at the time). Then we
increased the commission rate by a factor of about six, so that if the salesman brought in an order at a price
near list, then his commission was somewhat higher than before. If he started discounting, his commission
dropped sharply. We gave broad discretion to sales management to approve discounts.
"We still had sales guys claiming that a sale was massively strategic and had to be sold at or below cost,
and that they needed their commission anyway to justify the effort. In some cases we approved this, but
mostly we said that if it's strategic then you'll get your commission on the follow-on sales. While salesmen
have a strong preference for immediate cash, they will act so as to maximize income over time, and will
think and act strategically if financially rewarded for such.
"The results were that our margins increased significantly. Revenues were increasing too. It's hard to
attribute the revenue gain to the new commission plan, given the number of other variables, but I like to
think that it was a factor. Now our salesmen spent more of their creative energies devising ways to sell our
customers on the value of our products and company, instead of conspiring with sales management as to
the best tale of woe to present to marketing at the Monday pricing meeting.
"The Monday meetings became shorter and more pleasant, focused on truly creative ways to make each
sale. There certainly were steep discounts given in some cases, but they reflected the competitive situation
and future sales potential at each account much more accurately."
(Source: private correspondence, quotation permission received)
McAfee: Introduction to Economic Analysis, http://www.introecon.com, July 24, 2006
7-285
The agent won't accept employment unless u*>uo, the reservation utility. The principal
can minimize the cost of employing the agent by setting the salary such that u*=uo,
which results in
y = u0 -/2s2a+s5a2
Observe that the salary has to be higher, the greater is the risk 62. That is, the principal
has to cover the cost of risk in the salary term.
7.5.2 Cost of Providing Incentives
The principal obtains profits which are the remainder of the value after paying the
agent, and minus the salary:
T = (1-s)x - y
= (i-s)sa-(u0 -/2s2a+sla2)
= sa -U W-1/2a-s26
Note that the principal gets the entire output x = sa minus all the costs - the reservation
utility of the agent uo, the cost of providing effort, and the risk cost on the agent. That
is, the principal obtains the full gains from trade - the value of production minus the
total cost of production. However, the fact that the principal obtains the full gains from
trade doesn't mean the principal induces the agent to work extremely hard, because
there is no mechanism for the principal to induce the agent to work hard without
imposing more risk on the agent, and this risk is costly to the principal. Agents are
induced to work hard by tying their pay to their performance, and such a link
necessarily imposes risk on the agent, and risk is costly.10o
We take the principal to be risk neutral. This is reasonable when the principal is
"economically large" relative to the agent, so that the risks faced by the agent are small
to the principal. For example, the risks associated with any one car are small to a car
rental company. The principal who maximizes expected profits chooses s to maximize c,
which yields
S=1- -6
a
This formula is interesting for several reasons. First, if the agent is neutral to risk,
which means 2%=o, then s is 1. That is, the agent gets ioo% of the marginal return to
effort, and the principal just collects a lump sum. This is reminiscent of some tenancy
contracts used by landlords and peasants; the peasant paid a lump sum for the right to
farm the land and then kept all of the crops grown. Since these peasants were unlikely
to be risk neutral, while the landlord was relatively neutral to risk, such a contract was
100 There is a technical requirement that the principal's return ac must be positive, for otherwise the
principal would rather not contract at all. This amounts to an assumption that uo is not too large.
Moreover, if s comes out less than zero, the model falls apart, and in this case, the actual solution is s=o.
McAfee: Introduction to Economic Analysis, http://www.introecon.com, July 24, 2006
7.286
unlikely to be optimal. The contract with s=1 is known as "selling the agency" since the
principal sells the agency to the agent for a lump sum payment. (Here, y will generally
be negative - the principal gets a payment rather than paying a salary.) The more
common contract, however, had the landowner and the tenant farmer share the
proceeds of farming, which gives rise to the name sharecropper.
Second, more risk or more risk aversion on the part of the agent decreases the share of
the proceeds accruing to the agent. Thus, when the cost of risk or the amount of risk is
high, the best contract imposes less risk on the agent. Total output sa falls as the risk
costs rise.
Third, more able agents (higher a) get higher commissions. That is, the principal
imposes more risk on the more able agent because the returns to imposition of risk - in
the form of higher output - are greater, and thus worth the cost in terms of added risk.
Most real estate agencies operate on a mix of salary and commission, with commissions
paid to agents averaging about 50%. The agency RE/MAX, however, pays commissions
close to ioo%, collecting a fixed monthly fee that covers agency expenses from the
agents. RE/MAX claims that their formula is appropriate for better agents. The theory
developed suggests that more able agents should obtain higher commissions. But in
addition, RE/MAX's formula also tends to attract more able agents, since able agents
earn a higher wage under the high commission formula. (There is a potential downside
to the RE/MAX formula, that it discourages agency-wide cooperation.)
7.5.3 Selection of Agent
Consider what contracts attract what kinds of agents. For a fixed salary y and
commission s, the agent's utility, optimizing over x, is
u* y+1/2s2a-s2la2.
The agent's utility is increasing in a and decreasing in &. Thus, more able agents get
higher utility, and less risk averse agents get higher utility.
How do the terms of the contract affect the pool of applicants? Let us suppose two
contracts are offered, one with a salary y1 and commission si, the other with salary y2
and commission s2. We suppose y2 < y1 and s2 > s1. What kind of agent prefers contract
2, the high commission, low salary contract, over contract 1?
y 2 s9a -s21 y1+± W/2sa -s12
or, equivalently,
W/a(s2 -sjd-(s2 -si)Aa2 y1 - y2.-
Thus, agents with high ability a or low level of risk aversion 2A prefers the high
commission, low salary contract. A company that puts more of the compensation in the
form of commission tends to attract more able agents, and agents less averse to risk.
McAfee: Introduction to Economic Analysis, http://www.introecon.com, July 24, 2006
7.287
The former is a desirable feature of the incentive scheme, since more able agents
produce more. The latter, the attraction of less risk averse agents, may or may not be
desirable but is probably neutral overall.
One important consideration is that agents who overestimate their ability will react the
same as people who have higher ability. Thus, the contract equally attracts those with
high ability and those who overestimate their ability.
Agency theory provides a characterization of the cost of providing incentives. The
source of the cost is the link between incentives and risk. Incentives link pay and
performance; when performance is subject to random fluctuations, linking pay and
performance also links pay and the random fluctuations. Thus the provision of
incentives necessarily imposes risk on the agent, and if the agent is risk averse, this is
costly.
In addition, the extent to which pay is linked to performance will tend to affect the type
of agent who is willing to work for the principal. Thus, a principal must not only
consider the incentive to work hard created by the commission and salary structure, but
also the type of agent who would choose to accept such a contract.
7.5.4 Multi-tasking
Multi-tasking refers to performing several activities simultaneously. All of us multi-
task. We study while drinking a caffeinated beverage; we think about things in the
shower; we talk all too much on cell phones and eat French fries while driving. In the
context of employees, an individual employee will be assigned a variety of tasks and
responsibilities, and the employee must divide their time and efforts among the tasks.
Incentives provided to the employee must direct not only the total efforts of the
employee, but also the allocation of time and effort across activities. An important
aspect of multi-tasking is the interaction of incentives provided to employees, and the
effects of changes in one incentive on the behavior of the employee over many different
dimensions. In this section, we will establish conditions under which the problem of an
employer disaggregates, that is to say, the incentives on each individual task can be set
independently of the incentives applied to the others.
This section is relatively challenging and involves a number of pieces. To simplify the
presentation, some of the pieces are set aside as claims.
To begin the analysis, we consider a person who has n tasks or jobs. For convenience we
will index these activities with the natural numbers 1, 2, ... , n. The level of activity,
which may also be thought of as an action, in task i will be denoted by x1. It will prove
convenient to denote the vector of actions by x = (x1, ..., xv). We suppose the agent
bears a cost c(x) of undertaking the vector of actions x. We make four assumptions on
c:
* c is increasing in each xi,
* c has a continuous second derivative
* c is strictly convex, and
McAfee: Introduction to Economic Analysis, http://www.introecon.com, July 24, 2006
7.288
0 c is homogeneouslol of degree r.
For example, if there are two tasks (n=2), then all four of these assumptions are met by
the cost function c(x1 ,x2) = x +x2 + 1/2X1x2. This function is increasing in x1 and x2,
has continuous derivatives, is strictly convex (more about this below) and is
homogeneous of degree 2.
It is assumed that c is increasing to identify the activities as costly. Continuity of
derivatives is used for convenience. Convexity of c will insure that a solution to the first
order conditions is actually an optimum for the employee. Formally, it means that for
any vectors x # y and scalar a between zero and one (o _a i),
ac(x) + (1- a)c(y)>c(ax + (1- a)y).
One way of interpreting this requirement is that it is less costly to do the average of two
things than the average of the costs of the things. Intuitively, convexity requires that
doing a medium thing is less costly than the average of two extremes. This is plausible
when extremes tend to be very costly. It also means the set of vectors which cost less
than a fixed amount, {x I c(x) b}, is a convex set. Thus, if two points cost less than a
given budget, the line segment connecting them does, too. Convexity of the cost
function insures that the agent's optimization problem is concave, and thus that the first
order-conditions describe a maximum. When the inequality is strict for a satisfying o <
a < 1, we refer to convexity as strict convexity.
The assumption of homogeneity dictates that scale works in a particularly simple
manner. Scaling up activities increases costs at a fixed rate r. Homogeneity has very
strong implications that are probably unreasonable in many settings. Nevertheless,
homogeneity leads to an elegant and useful theory, as we shall see. Recall the definition
of a homogeneous function: c is homogeneous of degree r means that for any 2 > o,
c(2,x) = rc(x).
Claim: strict convexity implies that r > 1.
Proof of Claim: Fix any x and consider the two points x and 2,x. By convexity, for
o p, and lose if v < p. To win when v > p, and
lose if v < p, can be assured by bidding v. Essentially, there is no gain to bidding less
than your value, because your bid doesn't affect the price, only the likelihood of winning.
Bidding less than value causes the bidder to lose when the highest rival bid falls between
the bid and the value, which is a circumstance that the bidder would like to win.
Similarly, bidding more than value only creates a chance of winning when the price is
higher than the bidder's value, in which case the bidder would prefer to lose.
Thus, bidders in a Vickrey auction have a dominant strategy to bid their value. This
produces the same outcome as the English auction, however, because the payment made
is the second-highest value, which was the price in the English auction. Thus, the
Vickrey auction is a sealed-bid implementation of the English auction when bidders
have private values, producing the same outcome, which is that the highest value bidder
wins, but pays the second-highest value.
Because the Vickrey auction induces bidders to bid their value, it is said to be demand
revealing. Unlike the English auction, in which the bidding stops when the price
reaches the second-highest value and thus doesn't reveal the highest value, the Vickrey
auction reveals the highest value. In a controlled, laboratory setting, demand revelation
is useful, especially when the goal is to identify buyer values. Despite its theoretical
niceties, the Vickrey auction can be politically disastrous. Indeed, New Zealand sold
radio spectrum with the Vickrey auction on the basis of advice by a naive economist, and
the Vickrey auction created a political nightmare when a nationwide cellular license
received a high bid of $110 million, and a second-highest bid of $11 million. The
political problem was that the demand revelation showed that the government received
only about io% of the value of the license, making the public quite irate and dominating
news coverage at the time.107 Some smaller licenses sold for tenths of a percent of the
highest bid.
In a private values setting, the Vickrey auction, or the English auction, are much easier
on bidders than a regular sealed-bid auction, because of the dominant strategy. The
sealed-bid auction requires bidders to forecast their rivals' likely bids, and produces the
risks of either bidding more than necessary, or losing the bidding. Thus, the regular
sealed-bid auction has undesirable properties. Moreover, bidders in the sealed-bid
auction have an incentive to bribe the auctioneer to reveal the best bid by rivals, because
that is useful information in formulating a bid. Such (illegal) bribery occurs from time
to time in government contracting.
On the other hand, the regular sealed-bid auction has an advantage over the other two
that it makes price-fixing more difficult. A bidder can cheat on a conspiracy and not be
detected until after the current auction is complete.
Another disadvantage of the sealed-bid auction is that it is easier to make certain kinds
of bidding errors. In the U.S. PCS auctions, in which rights to use the radio spectrum
10 The Vickrey auction generally produces higher prices than regular sealed-bid auctions if bidders are
symmetric (share the same distribution of values), but is a poor choice of auction format when bidders are
not symmetric. Since the incumbent telephone company was expected to have a higher value than others,
the Vickrey auction was a poor choice for that reason as well.
McAfee: Introduction to Economic Analysis, http://www.introecon.com, July 24, 2006
7.300
for cellular phones was sold for around $20 billion, one bidder, intending to bid
$200,000, inadvertently bid $200,000,000. Such an error isn't possible in an English
auction, because prices rise at a measured pace. Such errors have little consequence in a
Vickrey auction, since getting the price wrong by an order of magnitude requires two
bidders to make such errors.
7.6.5 Winner's Curse
"I paid too much for it, but it's worth it."
-Sam Goldwyn
The analysis so far has been conducted under the restrictive assumption of private
values. In most contexts, bidders are not sure of the actual value of the item being sold,
and information held by others is relevant to the valuation of the item. If I estimate an
antique to be worth $5,000, but no one else is willing to bid more than $1,ooo, I might
revise my estimate of the value down. This revision leads bidders to learn from the
auction itself what the item is worth.
The early bidders in the sale of oil lease rights in the Gulf of Mexico (the outer
continental shelf) were often observed to pay more than the rights were worth. This
phenomenon came to be known as the winner's curse. The winner's curse is the fact
that the bidder who most overestimates the value of the object wins the bidding. Naive
bidders, who don't adjust for the winner's curse, will tend to lose money because they
only win the bidding when they've bid too high.
0.4
0.3
0.2
0.1
v-46 v-26 V v+26 v+46
Figure 7-9: Normally Distributed Estimates
Auctions, by their nature, select optimistic bidders. Consider the case of an oil lease
(right to drill for and pump oil) that has an unknown value v. Different bidders will
obtain different estimates of the value, and we may view these estimates as draws from a
normal distribution illustrated in Figure 7-9. The estimates are correct on average,
which is represented by the fact that the distribution is centered on the true value v.
Thus a randomly chosen bidder will have an estimate that is too high as often as it is too
McAfee: Introduction to Economic Analysis, http://www.introecon.com, July 24, 2006
7.301
low, and the average estimate of a randomly selected bidder will be correct. However,
the winner of an auction will tend to be bidder with the highest estimate, not a randomly
chosen bidder. The highest of five bidders will have an estimate that is too large 97% of
the time. The only way the highest estimate is not too large is if all the estimates are
below the true value. With ten bidders, the highest estimate is larger than the true value
with probability 99.9%, because the odds that all the estimates are less than the true
value is (1/2)10 = o.1%. This phenomenon - that auctions tend to select the bidder with
the highest estimate, and the highest estimate is larger than the true value most of the
time - is characteristic of the winner's curse.
A savvy bidder corrects for the winner's curse. Such a correction is actually quite
straightforward when a few facts are available, and here a simplified presentation is
given. Suppose there are n bidders for a common value good, and the bidders receive
normally distributed estimates that are correct on average. Let a be the standard
deviation of the estimates.1o8 Finally, suppose that no prior information is given about
the likely value of the good.
In this case, it is a straightforward matter to compute a correction for the winner's curse.
Because the winning bidder will generally be the bidder with the highest estimate of
value, the winner's curse correction should be the expected amount by which the highest
value exceeds the average value. This can be looked up in a table for the normal
distribution. The values are given for selected numbers n in Table 7-30. This shows, as
a function of the number of bidders, how much each bidder should reduce their estimate
of value to correct for the fact that auctions select optimistic bidders. The units are
standard deviations.
Table 7-30: Winner's Curse Correction
n 1 2 3 4 5 10 15
WCC (6) 0 .56 .85 1.03 1.16 1.54 1.74
n 20 25 50 100 500 1000 10,000
WCC (6) 1.87 1.97 2.25 2.51 3.04 3.24 3.85
For example, with one bidder, there is no correction, since it was supposed that the
estimates are right on average. With two bidders, the winner's curse correction is the
amount that the higher of two will be above the mean, which turns out to be 0.566, a
little more than half a standard deviation. This is the amount which should be
subtracted from the estimate to insure that, when the bidder wins, the estimated value is
on average correct. With four bidders, the highest is a bit over a whole standard
deviation. As is apparent from the table, the winner's curse correction increases
relatively slowly after ten or fifteen bidders. With a million bidders, it is 4.86ca.
108 The standard deviation is a measure of the dispersion of the distribution, and is the square root of the
average of the square of the difference of the random value and its mean. The estimates are also assumed
to be independently distributed around the true value. Note that estimating the mean adds an additional
layer of complexity.
McAfee: Introduction to Economic Analysis, http://www.introecon.com, July 24, 2006
7.302
The standard deviation a measures how much randomness or noise there is in the
estimates. It is a measure of the average difference between the true value and the
estimated value, and thus the average level of error. Oil companies know from their
history of estimation how much error arises in the company estimates. Thus, they can
correct their estimates to account for the winner's curse using their historical
inaccuracies.
Bidders who are imperfectly informed about the value of an item for sale are subject to
losses arising from the way auctions select the winning bidder. The winning bidder is
usually the bidder with the highest estimate, and that estimate is too high on average.
The difference between the highest estimate and the average estimate is known as the
winner's curse correction. The size of the winner's curse correction is larger the more
bidders there are but tends to grow slowly beyond a dozen or so bidders.
If the bidders have the same information on a common value item, they will generally
not earn profits on it. Indeed, there is a general principle that it is the privacy of
information, rather than the accuracy of information, that leads to profits. Bidders earn
profits on the information that they hold that is not available to others. Information
held by others will be built into the bid price and therefore not lead to profits.
7.6.6 Linkage
The U.S. Department of the Interior, when selling off-shore oil leases, not only takes an
upfront payment (the winning bid) but also takes 1/6 of the oil that is eventually
pumped. Such a royalty scheme links the payment made to the outcome, and in a way,
shares risk, since the payment is higher when there is more oil. Similarly, a book
contract provides an author with an upfront payment and a royalty. Many U.S.
Department of Defense purchases of major weapons systems involve cost-sharing,
where the payments made pick up a portion of the cost. Purchases of ships, for example,
generally involve 50% to 70% cost sharing, which means the DOD pays a portion of cost
overruns. The contract for U.S. television broadcast rights for the summer Olympics in
Seoul, South Korea, involved payments that depended on the size of the U.S. audience.
Royalties, cost-sharing and contingent payments generally link the actual payment to
the actual value, which is unknown at the time of the auction. Linkage shares risk,
which is a topic already considered in Section 7.5. But linkage does something else, too.
Linkage reduces the importance of estimates in the auction, replacing the estimates with
actual values. That is, the price a bidder pays for an object, when fully linked to the true
value, is just the true value. Thus, linkage reduces the importance of estimation in the
auction by taking the price out of the bidder's hands, at least partially.
The linkage principle1O9 states that, in auctions where bidders are buyers, the expected
price rises the more the price is linked to the actual value. (In a parallel fashion, the
expected price in an auction where bidders are selling falls.) Thus, linking price to value
generally improves the performance of auctions. While this is a mathematically deep
result, an extreme case is straightforward to understand. Suppose the government is
purchasing by auction a contract for delivery of 10,000 gallons of gasoline each week for
109 The linkage principle, and much of modern auction theory, was developed by Paul Milgrom (1948 - ).
McAfee: Introduction to Economic Analysis, http://www.introecon.com, July 24, 2006
7.303
the next year. Suppliers face risk in the form of gasoline prices; if the government buys
at a fixed price, the suppliers' bids will build in a cushion to compensate for the risk, and
for the winner's curse. In addition, because their estimates of future oil prices will
generally vary, they will earn profits based on their private information about the value.
In contrast, if the government buys only delivery and then pays for the cost of the
gasoline, whatever it might be, any profits that the bidders earned based on their ability
to estimate gasoline prices evaporates. The overall profit level of bidders falls, and the
overall cost of the gasoline supply can fall. Of course, paying the cost of the gasoline
reduces the incentive of the supplier to shop around for the best price, and that agency
incentive effect must be balanced against the reduction in bidder profits from the
auction to select a supplier.
7.6.7 Auction Design
We saw above that the English auction tends to reduce regret relative to sealed-bid
auctions, and that the linkage principle suggests tying payments to value where possible.
These are examples of auction design, in which auctions are designed to satisfy
objectives of the auction designer. Proper auction design should match the rules of the
auction to the circumstances of the bidders and the goal of the seller. Some of the
principles of auction design include:
" Impose an appropriate reserve price or minimum bid
" Use ascending price (English) auctions rather than sealed-bid
" Reveal information about the value of the item
" Conceal information about the extent of competition
" Handicap bidders with a known advantage
However, many of these precepts change if the seller is facing a cartel. For example, it is
easier for bidders to collude in a sealed-bid auction than in an English auction; and
reserve prices should be made relatively high.
Reserve prices (minimum bid) have several effects. They tend to force marginal bidders
to bid a bit higher, which increases bids of all bidders, reducing bidder profits.
However, reserve prices also lead to a failure to sell on occasion, and the optimal reserve
trades off this failure to sell against the higher prices. In addition, reserve prices may
reduce the incentive of bidders to investigate the sale, reducing participation, which is
an additional negative consideration for a high reserve price.
Ascending price auctions like the English auction have several advantages. Such
auctions reduce the complexity of the bidder's problem, because bidder's can stretch
their calculations out over time, and because bidders can react to the behavior of others
and not plan for every contingency in advance. In addition, because bidders in an
English auction can see the behavior of others, there is a linkage created - the price paid
by the winning bidder is influenced not just by that bidder's information but also by the
information held by others, tending to drive up the price, which is an advantage for the
seller.
One caveat to the selection of the English auction is that risk aversion doesn't affect the
outcome in the private values case. In contrast, in a sealed-bid auction, risk aversion
McAfee: Introduction to Economic Analysis, http://www.introecon.com, July 24, 2006
7.304
works to the advantage of the seller, because bidders bid a little bit higher than they
would have otherwise, to reduce the risk of losing. Thus, in the private values case, risk
averse bidders will bid higher in the sealed-bid auction than in the English auction.
When considering the revelation of information, there is always an issue of lying and
misleading. In the long-run, lying and misleading is found out, and thus the standard
approach is to ignore the possibility of lying. Making misleading statements is, in the
long-run, the same thing as silence, since those who repeatedly lie or mislead are
eventually discovered, and then not believed. Thus, in the long-run, a repeat seller has a
choice of being truthful or silent. Because of the linkage principle, the policy of
revealing truthful information about the value of the good for sale dominates the policy
of concealing information, because the revelation of information links the payment to
the actual outcome.
In contrast, revealing information about the extent of competition may not increase the
prices. Consider the case where occasionally there are three bidders, and sometimes
only one. If the extent of competition is concealed, bidders will bid without knowing the
extent of competition. If the bidders are risk neutral, it turns out that the revelation
doesn't matter and the outcomes are the same on average. If, in contrast, bidders are
risk averse, the concealment of information tends to increase the bid prices, because the
risk created by the uncertainty about the extent of competition works to the advantage
of the seller. Of course, it may be difficult to conceal the extent of competition in the
English auction, suggesting a sealed-bid auction instead.
Bidders with a large, known advantage have several deleterious effects. For example,
incumbent telephone companies generally are willing to pay more for spectrum in their
areas than outsiders are. Advantaged bidders discourage participation of others, since
the others are likely to lose. This can result in a bidder with an advantage facing no
competition and picking up the good cheaply. Second, rivals don't present much
competition to the advantaged bidder, even if the rivals do participate. Consequently,
when a bidder has a large advantage over rivals, it is advantageous to handicap the
advantaged bidder, favoring the rivals. This handicapping encourages participation and
levels the playing field, forcing the advantaged bidder to bid more competitively to win.
A common means of favoring disadvantaged bidders is by the use of bidder credits. For
example, with a 20% bidder credit for disadvantaged bidders, a disadvantaged bidder
only has to pay 8o% of the face amount of the bid. This lets such a bidder bid 25% more
(since a $100 payment corresponds to a $125 bid) than they would have otherwise,
which makes the bidder a more formidable competitor. Generally, the ideal bidder
credit is less than the actual advantage of the advantaged bidder.
Auction design is an exciting development in applied industrial organization, in which
economic theory and experience is used to improve the performance of markets. The
U.S. Federal Communications Commissions auctions of spectrum, were the first major
instance of auction design in an important practical setting, and the auction design was
credited with increasing the revenue raised by the government substantially.
McAfee: Introduction to Economic Analysis, http://www.introecon.com, July 24, 2006
7-305
7.7 Antitrust
In somewhat archaic language, a trust was a group of firms acting in concert, which is
now known as a cartel. The antitrust laws made such trusts illegal, and were intended to
protect competition. In the United States, these laws are enforced by the Department of
Justice's Antitrust Division, and by the Federal Trade Commission. The United States
began passing laws during a time when some European nations were actually passing
laws forcing firms to join industry cartels. By and large, however, the rest of the world
has since copied the U.S. antitrust laws in one version or another.
7.7.1 Sherman Act
The Sherman Act, passed in 1890, is the first significant piece of antitrust legislation. It
has two main requirements:
Section 1. Trusts, etc., in restraint of trade illegal; penalty
Every contract, combination in the form of trust or otherwise, or conspiracy, in restraint of trade
or commerce among the several States, or with foreign nations, is declared to be illegal. Every
person who shall make any contract or engage in any combination or conspiracy hereby declared
to be illegal shall be deemed guilty of a felony, and, on conviction thereof, shall be punished by
fine not exceeding $10,000,000 if a corporation, or, if any other person, $350,000, or by
imprisonment not exceeding three years, or by both said punishments, in the discretion of the
court.
Section 2. Monopolizing trade a felony; penalty
Every person who shall monopolize, or attempt to monopolize, or combine or conspire with any
other person or persons, to monopolize any part of the trade or commerce among the several
States, or with foreign nations, shall be deemed guilty of a felony, and, on conviction thereof, shall
be punished by fine not exceeding $io,ooo,ooo if a corporation, or, if any other person,
$350,000, or by imprisonment not exceeding three years, or by both said punishments, in the
discretion of the court.110
The phrase "in restraint of trade" is challenging to interpret. Early enforcement of the
Sherman Act followed the "Peckham Rule," named for noted Justice Rufus Peckham,
which interpreted the Sherman Act to prohibit contracts that reduced output or raised
prices, while permitting contracts that would increase output or lower prices.
In one of the most famous antitrust cases ever, the United States sued Standard Oil,
which had monopolized the transportation of oil from Pennsylvania to the east coast
cities of the United States, in 1911.
The exact meaning of the Sherman Act had not been settled at the time of the Standard
Oil case. Indeed, Supreme Court Justice Edward White suggested that, because
contracts by their nature set the terms of trade and thus restrain trade to those terms
and Section 1 makes contracts restraining trade illegal, one could read the Sherman Act
to imply all contracts were illegal. Chief Justice White concluded that, since Congress
couldn't have intended to make all contracts illegal, the intent must have been to make
unreasonable contracts illegal, and therefore concluded that judicial discretion is
necessary in applying the antitrust laws. In addition, Chief Justice White noted that the
110 The current fines were instituted in 1974; the original fines were $5,o00, with a maximum
imprisonment of one year. The Sherman Act is 15 U.S.C. § 1.
McAfee: Introduction to Economic Analysis, http://www.introecon.com, July 24, 2006
7.306
act makes monopolizing illegal, but doesn't make having a monopoly illegal. Thus,
Chief Justice White interpreted the act to prohibit certain acts leading to monopoly, but
not monopoly itself.
The legality of monopoly was further clarified through a series of cases, starting with the
1945 Alcoa case, in which the United States sued to break up the aluminum monopoly
Alcoa. The modern approach involves a two-part test. First, does the firm have
monopoly power in a market? If not, no monopolization has occurred and there is no
issue for the court. Second, if so, did the firm use illegal tactics to extend or maintain
that monopoly power? In the language of a later decision, did the firm engage in "the
willful acquisition or maintenance of that power as distinguished from growth or
development as a consequence of superior product, business acumen or historic
accident?" (U.S. v. Grinnell, 1966.)
There are several important points that are widely misunderstood and even misreported
in the press. First, the Sherman Act does not make having a monopoly illegal. Indeed,
three legal ways of obtaining a monopoly - a better product, running a better business,
or luck - are spelled out in one decision. It is illegal to leverage that existing monopoly
into new products or services, or to engage in anticompetitive tactics to maintain the
monopoly. Moreover, you must have monopoly power currently to be found guilty of
illegal tactics.
When the Department of Justice sued Microsoft over the incorporation of the browser
into the operating system and other acts (including contracts with manufacturers
prohibiting the installation of Netscape), the allegation was not that Windows was an
illegal monopoly. The DOJ alleged Microsoft was trying to use its Windows monopoly
to monopolize another market, the internet browser market. Microsoft's defense was
two-fold. First, it claimed not to be a monopoly, citing the 5% share of Apple. (Linux
had a negligible share at the time.) Second, it alleged a browser was not a separate
market but an integrated product necessary for the functioning of the operating system.
This defense follows the standard "two-part test."
Microsoft's defense brings up the question of "what is a monopoly?" The simple answer
to this question depends on whether there are good substitutes in the minds of
consumers - will they substitute to an alternate product in the event of some bad
behavior by the seller? By this test, Microsoft had an operating system monopoly in
spite of the fact that there was a rival, because for most consumers, Microsoft could
increase the price, tie the browser and MP3 player to the operating system, or even
disable Word Perfect, and the consumers would not switch to the competing operating
system. However, Microsoft's second defense, that the browser wasn't a separate
market, is a much more challenging defense to assess.
The Sherman Act provides criminal penalties, which are commonly applied in price-
fixing cases, that is, when groups of firms join together and collude to raise prices.
Seven executives of General Electric and Westinghouse, who colluded in the late 1950s
to set the prices of electrical turbines, spent several years in jail each, and there was over
$10o million in fines. In addition, Archer Daniels Midland executives went to jail after
their 1996 conviction for fixing the price of lysine, which approximately doubled the
McAfee: Introduction to Economic Analysis, http://www.introecon.com, July 24, 2006
7.307
price of this common additive to animal feed. When highway contractors are convicted
of bid-rigging, generally the conviction is under the Sherman Act, for monopolizing their
market.
7.7.2 Clayton Act
Critics of the Sherman Act, including famous "trust-buster" and President Teddy
Roosevelt, felt the ambiguity of the Sherman Act was an impediment to its use and that
the United States needed a more detailed law setting out a list of illegal activities. The
Clayton Act, 15 U.S.C. §§ 12-27, was passed in 1914 and it adds detail to the Sherman
Act. The same year, the FTC Act was passed, creating the Federal Trade Commission,
which has authority to enforce the Clayton Act, as well as engage in other consumer
protection activities.
The Clayton Act does not have criminal penalties, but does allow for monetary penalties
that are three times as large as the damage created by the illegal behavior.
Consequently, a firm, motivated by the possibility of obtaining a large damage award,
may sue another firm for infringement of the Clayton Act. A plaintiff must be directly
harmed to bring such a suit. Thus, customers who paid higher prices, or firms driven
out of business by exclusionary practices are permitted to sue under the Clayton Act.
When Archer Daniels Midland raised the price of lysine, pork producers who bought
lysine would have standing to sue, while final pork consumers who paid higher prices
for pork, but who didn't directly buy lysine, would not.
Highlights of the Clayton Act include:
" Section 2, which prohibits price discrimination that would lessen
competition,
" Section 3, which prohibits exclusionary practices that lessen competition,
such as tying, exclusive dealing and predatory pricing,
" Section 7, which prohibits share acquisition or merger that would lessen
competition or create a monopoly
The language "lessen competition" is generally understood to mean that a significant
price increase becomes possible - that is, competition has been harmed if the firms in
the industry can successfully increase prices.
Section 2 is also known as 'Robinson-Patman' because of a 1936 amendment by that
name. It prohibits price discrimination that lessens competition. Thus, price
discrimination to final consumers is legal under the Clayton Act; the only way price
discrimination can lessen competition is if one charges different prices to different
businesses. The logic of the law was articulated in the 1948 Morton Salt decision, which
concluded that lower prices to large chain stores gave an advantage to those stores, thus
injuring competition in the grocery market. The discounts in that case were not cost-
based, and it is permissible to charge different prices based on costs.
Section 3 rules out practices that lessen competition. A manufacturer who also offers
service for the goods it sells may be prohibited from favoring its own service
organization. Generally manufacturers may not require the use of the manufacturer's
McAfee: Introduction to Economic Analysis, http://www.introecon.com, July 24, 2006
7.308
own service. Moreover, an automobile manufacturer can't require the use of
replacement parts made by the manufacturer, and many car manufacturers have lost
lawsuits on this basis. In an entertaining example, Mercedes prohibited Mercedes
dealers from buying Bosch parts directly from Bosch, even though Mercedes itself was
selling Bosch parts to the dealers. This practice was ruled illegal because the quality of
the parts was the same as Mercedes (indeed, identical), so Mercedes' action lessened
competition.
Predatory pricing involves pricing below cost in order to drive a rival out of business. It
is relatively difficult for a firm to engage in predation, simply because it only makes
sense if, once the rival is eliminated, the predatory firm can then increase its prices and
recoup the losses incurred. The problem is that once the prices go up, entry becomes
attractive; what keeps other potential entrants away? One answer is reputation: a
reputation for a willingness to lose money in order to dominate market could deter
potential entrants. Like various rare diseases that happen more often on TV than in the
real world (e.g. Tourette's syndrome), predatory pricing probably happens more often in
textbooks than in the real world."
The Federal Trade Commission also has authority to regulate mergers that would lessen
competition. As a practical matter, the Department of Justice and the Federal Trade
Commission divide responsibility for evaluating mergers. In addition, other agencies
may also have jurisdiction over mergers and business tactics. The Department of
Defense has oversight of defense contractors, using a threat of "we're your only
customer." The Federal Communications Commission has statutory authority over
telephone and television companies. The Federal Reserve Bank has authority over
national and most other banks.
Most states have antitrust laws as well, and can challenge mergers that would affect
commerce in the state. In addition, attorneys general of many states may join the
Department of Justice or the Federal Trade Commission is suing to block a merger or in
other antitrust actions, or sue independently. For example, many states joined the
Department of Justice in its lawsuit against Microsoft. Forty-two states jointly sued the
major record companies over their "minimum advertised prices" policies, which the
states argued resulted in higher compact disc prices. The "MAP" case settlement
resulted in a modest payment to compact disc purchasers. The Federal Trade
Commission had earlier extracted an agreement to stop the practice.
7.7.3 Price-Fixing
Price-fixing, which is called bid-rigging in a bidding context, involves a group of firms
agreeing to increase the prices they charge and restrict competition against each other.
The most famous example of price-fixing is probably the "Great Electrical Conspiracy"
in which GE and Westinghouse (and a smaller firm, Allis-Chalmers and many others)
fixed the prices of turbines used for electricity generation. Generally these turbines
were the subject of competitive (or in this case not-so-competitive) bidding, and one
way that the companies set the prices was to have a designated winner for each bidding
situation, and using a price book to provide identical bids by all companies. An amusing
m1 Economists have argued that American Tobacco, Standard Oil and A.T. & T. each engaged in predation
in their respective industries.
McAfee: Introduction to Economic Analysis, http://www.introecon.com, July 24, 2006
7.309
element of the price-fixing scheme was the means by which the companies identified the
winner in any given competition: it used the phase of the moon. The phase of the moon
determined the winner and each company knew what to bid based on the phase of the
moon. Executives from the companies met often to discuss terms of the price-fixing
arrangement, and the Department of Justice acquired a great deal of physical evidence
in the process of preparing its 1960 case. Seven executives went to jail and hundreds of
millions of dollars in fines were paid.
Most convicted price-fixers are from small firms. The turbine conspiracy and the Archer
Daniels Midland lysine conspiracy are unusual. (There is evidence that large vitamins
manufacturers conspired in fixing the price of vitamins in many nations of the world.)
Far more common conspiracies involve highway and street construction firms,
electricians, water and sewer construction companies or other "owner operated"
businesses. Price-fixing seems most common when owners are also managers and there
are a small number of competitors in a given region.
As a theoretical matter, it should be difficult for a large firm to motivate a manager to
engage in price-fixing. The problem is that the firm can't write a contract promising the
manager extraordinary returns for successfully fixing prices because such a contract
itself would be evidence and moreover implicate higher management. Indeed, Archer
Daniels Midland executives paid personal fines of $350,000 as well as each serving two
years in jail. Thus, it is difficult to offer a substantial portion of the rewards of price-
fixing to managers, in exchange for the personal risks the managers would face from
engaging in price-fixing. Most of the gains of price-fixing accrue to shareholders of large
companies, while large risks and costs fall on executives. In contrast, for smaller
businesses in which the owner is the manager, the risks and rewards are borne by the
same person, and thus the personal risk more likely to be justified by the personal
return.
We developed earlier a simple theory of cooperation, in which the grim trigger strategy
was used to induce cooperation. Let us apply that theory to price-fixing. Suppose that
there are n firms, and that they share the monopoly profits 7Cm equally if they collude. If
one firm cheats, that firm can obtain the entire monopoly profits until the others react.
This is clearly the most the firm could get from cheating. Once the others react, the
collusion breaks down and the firms earn zero profits (the competitive level) from then
on. The cartel is feasible if 1/n of the monopoly profits forever is better than the whole
monopoly profits for a short period of time. Thus, cooperation is sustainable if:
T I m _ m ( 1 + 6 + 2 + ) > m -
n(1-6) n
The left hand side gives the profits from cooperating - the present value of the 1/n share
of the monopoly profits. In contrast, if a firm chooses to cheat, it can take at most the
monopoly profits, but only temporarily. How many firms will this sustain? The
inequality simplifies to n s 1 Suppose the annual interest rate is 5% and the
1-6
reaction time is 1 week - that is, a firm that cheats on the cooperative agreement
McAfee: Introduction to Economic Analysis, http://www.introecon.com, July 24, 2006
7.310
sustains profits for a week, after which time prices fall to the competitive level. In this
case, 1-6 is a week's worth of interest (6 is the value of money received in a week) and
therefore 6=0.95 52=.999014. According to standard theory, the industry with a
week-long reaction time should be able to support cooperation with up to a thousand
firms!
There are a large variety of reasons why this theory fails to work very well empirically,
including that some people are actually honest and don't break the law, but we will focus
on one game-theoretic reason here. The cooperative equilibrium is not the only
equilibrium, and there are good reasons to think that full cooperation is unlikely to
persist. The problem is the prisoner's dilemma itself: generally the first participant to
turn in the conspiracy can avoid jail. Thus, if one member of a cartel is uncertain
whether the other members of a price-fixing conspiracy are contacting the Department
of Justice, that member may race to the DOJ - the threat of one confession may cause
them all to confess in a hurry. A majority of the conspiracies that are prosecuted arise
because someone - a member who feels guilty, a disgruntled ex-spouse of a member, or
perhaps a member who thinks another member is suffering pangs of conscience - turns
them in. Lack of confidence in the other members creates a self-fulfilling prophecy.
Moreover, cartel members should lack confidence in the other cartel members who are,
after all, criminals.
On average, prosecuted conspiracies were about seven years old when they were caught.
Thus, there is about a 15% chance annually of a breakdown of a conspiracy, at least
among those that are eventually caught.
7.7.4 Mergers
The U.S. Department of Justice and the Federal Trade Commission share responsibility
for evaluating mergers. Firms with more than $50 million in assets are required under
the Hart-Scott-Rodino Act to file an intention to merge with the government. The
government then has a limited amount of time to either approve the merger or request
more information (called a second request). Once the firms have complied with the
second request, the government again has a limited amount of time before it either
approves the merger or sues to block it. The agencies themselves don't stop the merger,
but instead sue to block the merger, asking a federal judge to prevent the merger as a
violation of one of the antitrust laws. Mergers are distinct from other violations,
because they have not yet occurred at the time the lawsuit is brought, so there is no
threat of damages or criminal penalties; the only potential penalty imposed on the
merging parties is that the proposed merger may be blocked.
Many proposed mergers result in settlements. As part of the settlement associated with
GE's purchase of RCA in 1986, a small appliance division was sold to Black & Decker,
thereby maintaining competition in the small kitchen appliance market. In the 1999
merger of oil companies Exxon and Mobil, a California refinery, shares in oil pipelines
connecting the gulf with the northeast, and thousands of gas stations were sold to other
companies. The 1996 merger of Kimberley-Clark and Scott Paper would have resulted
in a single company with over 50% of the facial tissue and baby wipes markets, and in
both cases divestitures of production capacity and the "Scotties" brand name preserved
McAfee: Introduction to Economic Analysis, http://www.introecon.com, July 24, 2006
7.311
competition in the markets. Large bank mergers, oil company mergers and other large
companies usually present some competitive concerns, and the majority of these cases
are solved by divestiture of business units to preserve competition.
A horizontal merger is a merger of competitors, such as Exxon and Mobil or two banks
located in the same city. In contrast, a vertical merger is a merger between an input
supplier and input buyer. The attempt by book retailer Barnes and Noble to purchase
the intermediary Ingram, a company that buys books from publishers and sells to
retailers but didn't directly sell to the public, would have resulted in a vertical merger.
Similarly, Disney is a company that sells programs to television stations (among other
activities), so its purchase of TV network ABC was a vertical merger. The AOL--Time
Warner merger involved several vertical relationships. For example, Time Warner is a
large cable company, and cable represents a way for AOL to offer broadband services.
In addition, Time Warner is a content provider, and AOL delivers content to internet
subscribers.
Vertical mergers raise two related problems:foreclosure and raising rivals' costs.
Foreclosure refers to denying access to necessary inputs. Thus, the AOL--Time Warner
merger threatened rivals to AOL internet service (like EarthLink) with an inability to
offer broadband services to consumers with Time Warner cable. This potentially injures
competition in the internet service market, forcing Time Warner customers to use AOL.
In addition, by bundling Time Warner content and AOL internet service, users could be
forced to purchase AOL internet service in order to have access to Time Warner content.
Both of these threaten foreclosure of rivals, and both were resolved to the government's
satisfaction by promises that the merged firm would offer equal access to rivals.
Raising rivals' costs is a softer version of foreclosure. Rather than deny access to
content, AOL--Time Warner could instead make the content available under
disadvantageous terms. For example, American Airlines developed the Sabre
computerized reservation system, which was used by about 40% of travel agents. This
system charged airlines, rather than travel agents, for bookings. Consequently,
American Airlines had a mechanism for increasing the costs of its rivals, by increasing
the price of bookings on the Sabre system. The advantage to American was not just
increased revenue of the Sabre system but also the hobbling of airline rivals. Similarly,
banks offer free use of their own automated teller machines (ATMs), but charge the
customers of other banks. Such charges raise the costs of customers of other banks,
thus making other banks less attractive, and hence providing an advantage in the
competition for bank customers.
The Department of Justice and the Federal Trade Commission periodically issue
horizontal merger guidelines, which set out how mergers will be evaluated. This is a
three step procedure for each product that the merging companies have in common.
The procedure starts by identifying product markets. To identify a product market, start
with a product or products produced by both companies. Then ask if the merged parties
can profitably raise price by a small but significant and non-transitory increase in
price, also known as a "SSNIP," pronounced 'snip.' A SSNIP is often taken to be a 5%
price increase, which must prevail for several years. If the companies can profitably
McAfee: Introduction to Economic Analysis, http://www.introecon.com, July 24, 2006
7.312
increase price by a SSNIP, then they are judged to have monopoly power and consumers
will be directly harmed by the merger. (This is known as a unilateral effect, because the
merging parties can increase price unilaterally after the merger is consummated.) If
they can't increase prices, then an additional product has to be added to the group;
generally the best substitute is added. Ask whether a hypothetical monopoly seller of
these three products can profitably raise price. If so, an antitrust market has been
identified; if not, yet another substitute product must be added. The process stops
adding products when enough substitutes have been identified which, if controlled by a
hypothetical monopoly, would have their prices significantly increased.
The logic of product market definition is that, if a monopoly wouldn't increase price in a
meaningful way, that there is no threat to consumers - any price increase won't be large
or won't last. The market is defined by the smallest set of products for which consumers
can be harmed. The test is also known as the hypothetical monopoly test.
The second step is to identify a geographic market. The process starts with an area in
which both companies sell, and asks if the merged company has an incentive to increase
price by a SSNIP. If so, that geographic area is a geographic market. If not, it is because
of buyers substituting outside the area to buy cheaply, and the area must be expanded.
For example, owning all the gas stations on a corner doesn't let one increase price
profitably because an increase in price leads to substitution to stations a few blocks
away. If one company owned all the stations in a half mile radius, would it be profitable
to increase price? Probably not, as there would still be significant substitution to more
distant stations. Suppose, instead, that one owned all the stations for a 15 mile radius.
Then an increase in price in the center of the area is not going to be thwarted by too
much substitution outside the area, and the likely outcome is that prices would be
increased by such a hypothetical monopoly. In this case, a geographic market has been
identified. Again, parallel to the product market definition, a geographic market is the
smallest area in which competitive concerns would be raised by a hypothetical
monopoly. In any smaller area, attempts to increase price are defeated by substitution
to sellers outside the area.
The product and geographic markets together are known as a relevant antitrust market,
relevant for the purposes of analyzing the merger.
The third and last step of the procedure is to identify the level of concentration in each
relevant antitrust market. The Hirschman-Herfindahl index, or HHI, is used for this
purpose. The HHI is the sum of the squared market shares of the firms in the relevant
antitrust market, and is justified because it measures the price - cost margin in the
Cournot model. Generally in practice the shares in percent are used, which makes the
scale range from 0 to 10,000. For example, if one firm has 40%, one 30%, one 20% and
the remaining firm io%, the H HI is
402 + 302 + 202 + 102 = 3,000.
Usually, anything over i8oo is considered "very concentrated," and anything over 1000
is "concentrated."
McAfee: Introduction to Economic Analysis, http://www.introecon.com, July 24, 2006
7.313
Suppose firms with shares x and y merge, and nothing in the industry changes besides
the combining of those shares. Then the HHI goes up by (x + y)2 - x2 -y2 = 2xy. This
is referred to as the change in the HHI. The merger guidelines suggest the government
will likely challenge mergers with (i) a change of 100 and a concentrated post-merger
HHI, or (ii) a change of 50 and a very concentrated post-merger HHI. It is more
accurate to understand the merger guidelines to say that the government likely won't
challenge unless either (i) or (ii) is met. Even if the post-merger HHI suggests a very
concentrated industry, the government is unlikely to challenge is the change in the HHI
is less than 50.
Several additional factors affect the government's decision. First, if the firms are
already engaging in price discrimination, the government may define quite small
geographic markets, and possibly as small as a single customer. Second, if one firm is
very small (less than a percent) and the other not too large (less than 35%) the merger
may escape scrutiny because the effect on competition is likely small. Third, if one firm
is going out of business, the merger may be allowed as a means of keeping the assets in
the industry. Such was the case with Greyhound's takeover of Trailways, a merger to
monopoly of the only intercity bus companies in the United States.
Antitrust originated in the United States and the United States remains the most
vigorous enforcer of antitrust laws. However, the European Union has recently taken a
more aggressive antitrust stance and in fact blocked mergers that obtained tentative
U.S. approval, such as General Electric and Honeywell.
Antitrust is, in some sense, the applied arm of oligopoly theory. Because real situations
are so complex, the application of oligopoly theory to antitrust analysis is often
challenging, and we have only scratched the surface of many of the more subtle issues of
law and economics in this text. For example, intellectual property, patents and
standards all have their own distinct antitrust issues.
McAfee: Introduction to Economic Analysis, http://www.introecon.com, July 24, 2006
7.314
8 Index
8.i List of Figures
FIGURE 2-1: THE DEMAND CURVE ...................................................................................................................................................................2-9
FIGURE 2-2: CONSUMER SURPLUS.................................................................................................................................................................2-11
FIGURE 2-3: AN INCREASE IN DEMAND .......................................................................................................................................................2-12
FIGURE 2-4: THE SUPPLY CURVE....................................................................................................................................................................2-14
FIGURE 2-5: SUPPLIER PROFITS......................................................................................................................................................................2-15
FIGURE 2-6: AN INCREASE IN SUPPLY ..........................................................................................................................................................2-16
FIGURE 2-7: MARKET DEMAND .......................................................................................................................................................................2-19
FIGURE 2-8: EQUILIBRATION...........................................................................................................................................................................2-21
FIGURE 2-9: AN INCREASE IN DEMAND .......................................................................................................................................................2-23
FIGURE 2-10: AN INCREASE IN SUPPLY ........................................................................................................................................................2-24
FIGURE 2-11: PRICE OF STORAGE....................................................................................................................................................................2-25
FIGURE 2-12: ELASTICITIES FOR LINEAR DEMAND..................................................................................................................................2-29
FIGURE 2-13: THE PRODUCTION POSSIBILITIES FRONTIER .................................................................................................................2-33
FIGURE 2-14: TWO PRODUCTION POSSIBILITIES FRONTIERS ..............................................................................................................2-34
FIGURE 2-15: JOINT PPF .....................................................................................................................................................................................2-35
FIGURE 3-1: US RESIDENT POPULATION......................................................................................................................................................3-41
FIGURE 3-2: US URBAN AND WHITE POPULATION...................................................................................................................................3-42
FIGURE 3-3: POPULATION PROPORTIONS BY AGE GROUP ....................................................................................................................3-42
FIGURE 3-4: PROPORTION OF POPULATION UNDER AGE FIVE...........................................................................................................3-43
FIGURE 3-5: US LIFE EXPECTANCY AT BIRTH.............................................................................................................................................3-44
FIGURE 3-6: US IMMIGRANT POPULATION, IN PERCENT, BY CONTINENT OF ORIGIN................................................................3-44
FIGURE 3-7: NATIONAL ORIGIN OF IMMIGRANTS, 1900-2000 ..............................................................................................................3-45
FIGURE 3-8: MALE MARITAL STATUS (PERCENTAGES)...........................................................................................................................3-45
FIGURE 3-9: FEMALE MARITAL STATUS (PERCENT) ................................................................................................................................3-46
FIGURE 3-10: PERCENT OF BIRTHS TO UNWED MOTHERS....................................................................................................................3-46
FIGURE 3-11: PERCENT OF BIRTHS TO WOMEN AGE 19 OR LESS .........................................................................................................3-47
FIGURE 3-12: EDUCATIONAL ATTAINMENT IN YEARS (PERCENT OF POPULATION) ....................................................................3-47
FIGURE 3-13: GRADUATION RATES ................................................................................................................................................................3-48
FIGURE 3-14: HOUSEHOLD OCCUPANCY......................................................................................................................................................3-49
FIGURE 3-15: PROPORTION OF HOUSEHOLDS BY TYPE ..........................................................................................................................3-50
FIGURE 3-16: PERCENTAGE OF INCARCERATED RESIDENTS................................................................................................................3-50
FIGURE 3-17: INCOME SHARES FOR THREE QUINTILES .........................................................................................................................3-51
FIGURE 3-18: FAMILY INCOME.........................................................................................................................................................................3-51
FIGURE 3-19: FAMILY INCOME, CUMULATIVE PERCENTAGE CHANGE .............................................................................................3-52
FIGURE 3-20: BLACK FAMILY INCOME AS A PERCENTAGE OF WHITE INCOME .............................................................................3-53
FIGURE 3-21: CONSUMER PRICE INDEX (1982 = 100) ...............................................................................................................................3-53
FIGURE 3-22: CPI PERCENT CHANGES ..........................................................................................................................................................3-54
FIGURE 3-23: FOOD EXPENDITURE AS PERCENT OF INCOME, AND PROPORTION SPENT OUT................................................3-55
FIGURE 3-24: AFTER TAX INCOME SHARES ................................................................................................................................................3-55
FIGURE 3-25: OUTPUT, CONSUMPTION, INVESTMENT AND GOVERNMENT ...................................................................................3-57
FIGURE 3-26: MAJOR GDP COMPONENTS IN LOG SCALE .......................................................................................................................3-58
FIGURE 3-27: PER CAPITA INCOM E AND CONSUM PTION .....................................................................................................................3-58
FIGURE 3-28: CONSUMPTION, INVESTMENT AND GOVERNMENT (% GDP).....................................3-59
FIGURE 3-29: US AGRICULTURAL OUTPUT, 1982 CONSTANT DOLLARS ............................................................................................3-60
FIGURE 3-30: AGRICULTURAL OUTPUT, TOTAL AND PER WORKER (1982 $, LOG SCALE)...........................................................3-60
FIGURE 3-31: MAJOR NON-AGRICULTURAL SECTORS OF US ECONOMY, PERCENT OF GDP......................................................3-61
FIG U R E 3-32: A IR TRA V EL PER CA PITA ...... ...............................................................................................................................................3-62
FIGURE 3-33: ELECTRICITY PRODUCTION (M KWH)................................................................................................................................3-62
FIGURE 3-34: ENERGY USE (QUADRILLION BTUS) ...................................................................................................................................3-63
FIGURE 3-35: CARS PER THOUSAND POPULATION AND MILES DRIVEN PER CAPITA..................................................................3-63
FIGURE 3-36: PERCENTAGE OF POPULATION EMPLOYED (MILITARY & PRISONERS EXCLUDED) ..........................................3-64
FIGURE 3-37: LABOR FORCE PARTICIPATION RATES, ALL WOMEN AND MARRIED WOMEN....................................................3-64
FIGURE 3-38: DEFENSE AS A PERCENTAGE OF GDP..........................................................................................3-65
FIGURE 3-39: FEDERAL EXPENDITURES AND REVENUES (PERCENT OF GDP).......................................................3-66
FIGURE 3-40: FEDERAL, STATE & LOCAL AND TOTAL GOVERNMENT RECEIPTS (% GDP)..........................................3-66
FIGURE 3-41: FEDERAL, REGIONAL AND TOTAL EXPENDITURES AS A PERCENT OF GDP........................................3-67
FIGURE 3-42: FEDERAL EXPENDITURES, ON AND OFF BUDGET, PERCENT OF GDP................................................3-67
FIGURE 3-43: FEDERAL AND REGIONAL GOVERNMENT EMPLOYMENT (0OOS)......................................................3-68
FIGURE 3-44: MAJOR EXPENDITURES OF THE FEDERAL GOVERNMENT................................................................3-68
FIGURE 3-45: MAJOR TRANSFER PAYMENT S (% OF FEDERAL BUDGET)................................................................3-69
FIGURE 3-46: SOCIAL SECURITY REVENUE AND EXPENDITURE, $ MILLIONS .........................................................3-69
FIGURE 3-47: FEDERAL DEBT, TOTAL AND PERCENT OF GDP ...............................................................................3-70
FIGURE 3-48: FEDERAL SPENDING ON R&D, AS A PERCENT OF GDP......................................................................3-72
FIGURE 3-49: SOURCES OF FEDERAL GOVERNMENT REVENUE...........................................................................3-73
FIGURE 3-50: TOTAL IMPORT S AND EXPORTS AS A PROPORTION OF GDP.............................................................3-74
McAfee: Introduction to Economic Analysis, http://www.introecon.com, July 24, 2006
8-315
FIGURE 3-51: US TRADE IN GOODS AND SERVICES.............................................................................................. 3-74
FIGURE 3-52: INCOME AND PAYMENTS AS A PERCENT OF GDP............................................................................. 3-75
FIGURE 3-53: POSTWAR INDUSTRIAL PRODUCTION AND RECESSIONS.................................................................. 3-77
FIGURE 3-54: PERCENTAGE OF THE POPULATION EMPLOYED ............................................................................. 3-77
FIGURE 3-55: INDUSTRIAL FACTORY CAPACITY UTILITZATION (SOURCE: FRED) .................................................... 3-78
FIGURE 4-1: COBB-DOUGLAS ISOQUANTS...................................................................................................... 4-82
FIGURE 4-2: THE PRODUCTION FUNCTION..................................................................................................... 4-82
FIGURE 4-3: FIXED PROPORTIONS ................................................................................................................ 4-83
FIGURE 4-4: PERFECT SUBSTITUTES............................................................................................................. 4-84
FIGURE 4-5: PROFIT-MAXIMIZING LABOR INPUT............................................................................................ 4-86
FIGURE 4-6: TANGENCY AND ISOQUANTS ...................................................................................................... 4-94
FIGURE 4-7: SHORT-RUN SUPPLY................................................................................................................... 4-98
FIGURE 4-8: AVERAGE AND MARGINAL COSTS................................................................................................ 4-99
FIGURE 4-9: INCREASED PLANT SIZE ........................................................................................................... 4-100
FIGURE 4-10: LONG-RUN EQUILIBRIUM....................................................................................................... 4-105
FIGURE 4 11: 4-11: N AE ASHIFT.........DE M A N D........................................................................................................ 4-10610
FIGURE 4-12: RETURN TO LONG-RUN EQUILIBRIUM......................................................................................... 4-106
FIGURE 4-13: A DECREASE IN DEMAND........................................................................................................ 4-107
FIGURE 4-14: A BIG DECREASE IN DEMAND.................................................................................................. 4-108
FIGURE 4-15: A DECREASE IN SUPPLY ........................................................................................................... 4-109
FIGURE 4-16: EQUILIBRIUM WITH EXTERNAL SCALE ECONOMY ........................................................................ 4-110
FIGURE 4-17: DECREASE IN DEMAND ........................................................................................................... 4-111
FIGURE 4-18: LONG-RUN AFTER A DECREASE IN DEMAND................................................................................ 4-111
FIGURE 4-20: DRAM REVENUE CYCLE......................................................................................................... 4-113
FIGURE 4-21: INVESTMENT STRIKE PRICE GIVEN INTEREST RATE R IN PERCENT ................................................. 4-123
FIGURE 4-22 INVESTMENT STRIKE PRICE GIVEN INTEREST RATER IN PERCENT ................................................. 4-124
FIGURE 4-23: OPTIMAL SOLUTION FOR T..................................................................................................... 4-129
FIGURE 4-24: THE PORSCHE SPEEDSTER...................................................................................................... 4-130
FIGURE 4-25: $500 CONFEDERATE STATES BILL ............................................................................................ 4-134
FIGURE 4-26: WESTERN ELECTRIC MODEL 500 TELEPHONE.............................................................................. 4-135
FIGURE 4-27: PRICES OVER A CYCLE FOR SEASONAL COMMODITIES ................................................................... 4-137
FIGURE 4-28: LOG OF PRICE OF GOLD OVER TIME......................................................................................... 4-138
FIG U R E 5 1: U D-1:S T BU D G ET.....SET.................................................................................................................... 5-14114
FIGURE 5-2: EFFECT OF AN INCREASE IN PRICE ON THE BUDGET....................................................................... 5-142
FIGURE 5-3: AN INCREASE IN INCOME......................................................................................................... 5-142
FIGURE 5-4: UTILITY ISOQUANTS ................................................................................................................ 5-144
FIGURE 5-5: CONVEX PREFERENCES............................................................................................................ 5-145
FIGURE 5-6: GRAPHICAL UTILITY MAXIMIZATION............................................................................................ 5-147
FIGURE 5-7: "CONCAVE" PREFERENCES, PREFER BOUNDARIES.......................................................................... 5-148
FIGURE 5-8: ISOQUANTS FORA BLISS POINT................................................................................................ 5-150
FIGURE 5-9: SUBSTITUTION WITH AN INCREASE IN PRICE ................................................................................ 5-151
FIGURE 5-10: SUBSTITUTION EFFECT.......................................................................................................... 5-152
FIGURE 5-12: BACKWARD BENDING - INFERIOR GOOD ................................................................................... 5-156
FIGURE 5-13: INCOME AND SUBSTITUTION EFFECTS......................................................................................... 5-157
FIGURE 5-14: QUASILINEAR ISOQUANTS........................................................................................................ 5-159
FIGURE 5-16: HOUSE PRICE GRADIENT........................................................................................................ 5-166
FIGURE 5-17: BORROWING AND LENDING..................................................................................................... 5-170
FIGURE 5-18: INTEREST RATE CHANGE........................................................................................................ 5-17 1
FIGURE 5-19: INTEREST RATE INCREASE ON LENDERS.................................................................................... 5-172
FIGURE 5-20: DIFFERENT RATES FOR BORROWING AND LENDING..................................................................... 5-173
FIGURE 5-21: THE EFFECT OF A TRANSITORY INCOME INCREASE....................................................................... 5-173
FIGURE 5-22: EXPECTED UTILITY AND CERTAINTY EQUIVALENTS...................................................................... 5-175
FIGURE 5-23: THE EDGEWORTH BOX ........................................................................................................... 5-182
FIGURE 5-24: AN EFFICIENT POINT.............................................................................................................. 5-183
FIGURE 5-25: THE CONTRACT CURVE.......................................................................................................... 5-184
McAfee: Introduction to Economic Analysis, http://www.introecon.com, July 24, 2006 8-31 6
FIGURE 6-11: A QUOTA ......................................................................................................................................................................................6-212
FIGURE 6-12: A NEGATIVE EXTERNALITY..................................................................................................................................................6-214
FIGURE 6-13: EXTERNAL COSTS AND BENEFITS......................................................................................................................................6-215
FIGURE 6-14: THE PIGOUVIAN TAX..............................................................................................................................................................6-217
FIGURE 6-15: SO2 PERMIT PRICES.................................................................................................................................................................6-219
FIGURE 6-16: FISH POPULATION DYNAMICS ............................................................................................................................................6-222
FIGURE 6-17: FISH POPULATION DYNAMICS WITH FISHING..............................................................................................................6-223
FIGURE 6-18: FISH POPULATION DYNAMICS: EXTINCTION ................................................................................................................6-224
FIGURE 6-19: POSSIBILITY OF MULTIPLE EQUILIBRIA..........................................................................................................................6-225
FIGURE 6-20: BASIC MONOPOLY DIAGRAM ..............................................................................................................................................6-234
FIGURE 6-21: TWO-PART PRICING................................................................................................................................................................6-241
FIGURE 6-22: NATURAL MONOPOLY ...........................................................................................................................................................6-242
FIGURE 7-1 SEQUENTIAL BANK LOCATION (NYC PAYOFF LISTED FIRST).......................................................................................7-266
FIG U R E 7-2: SU B G A M E PE R FE C T IO N ...........................................................................................................................................................7-267
FIGURE 7-3: CAN'T AVOID ROCKY .................................................................................................................................................................7-267
FIGURE 7-4: EXPECTED PRICES IN SEARCH EQUILIBRIUM.................................................................................................................7-277
FIGURE 7-5: EXPECTED PRICES (S=PROPORTION OF SHOPPERS).....................................................................................................7-278
FIGURE 7-6: HOTELLING MODEL FOR BREAKFAST CEREALS .............................................................................................................7-279
FIGURE 7-7: SHARING THE HOTELLING MARKET...................................................................................................................................7-280
FIGURE 7-8: A SEGMENT OF THE CIRCLE MODEL...................................................................................................................................7-281
FIGURE 7-9: NORMALLY DISTRIBUTED ESTIMATES ..............................................................................................................................7-301
8.2 Index
A
ABC, 7-312
ADM, 4-79, 7-307, 7-308, 7-310
Agency, 3-72, 5-177, 6-218,7-283, 7-284,7-286, 7-
287,7-288, 7-291, 7-292, 7-293, 7-304
Akerlof, George, 6-245, 6-246
Alcoa, 7-307
Allocation, 1-1, 2-35, 2-37, 2-38, 5-182, 6-219, 7-
288
Ambiguity, 2-37, 6-211, 7-308
American Airlines, 6-239, 7-312
Analysis
Normative, 1-2
Positive, 1-2
Antitrust, 1-2, 7-271, 7-306, 7-309, 7-311, 7-313, 7-
314
Clayton Act, 7-308
FTC Act, 7-308
Market Power, 6-232, 7-273
Monopoly Power, 6-232, 6-236, 7-273, 7-274, 7-
307, 7-313
Peckham Rule, 7-306
Predatory Pricing, 7-308, 7-309
Robinson Patman, 7-308
Sherman Act, 7-306, 7-307, 7-308
SSNIP, 7-312, 7-313
Tying, 7-286, 7-304, 7-308
Vertical Merger, 7-312
AOL, 7-312
Arbitrage, 4-114, 4-115, 4-126, 4-136, 6-238, 6-240
Arbitrage Condition, 4-126, 4-136
Attitude, 7-284
Auction, 1-1, 7-295, 7-298, 7-299, 7-300, 7-301, 7-
302, 7-303,7-304, 7-305
Bid-Increment, 7-296, 7-299
Common Values, 7-295, 7-302, 7-303
English, 7-295, 7-296, 7-298, 7-299, 7-300, 7-
301, 7-304, 7-305
Linkage Principle, 7-303, 7-304, 7-305
Sealed-Bid, 7-296, 7-297, 7-298, 7-299, 7-300,
7-304, 7-305
Vickrey, 7-299, 7-300, 7-301
Autarky, 5-185, 5-186
Availability, 4-126
B
Bargaining, 5-180, 5-186, 6-221, 6-229, 6-246
Basketball, 4-79, 5-164, 5-179
Beer, 2-9, 2-12, 5-140, 5-150
Benefit Concentrated, 6-210
Bliss Point, 5-150
Boeing, 4-101
Bosch, 7-309
Brands, 4-127, 7-311
British Petroleum, 1-2
Budget Line, 5-141, 5-143, 5-146, 5-147, 5-149, 5-
151, 5-152, 5-154, 5-187
Budget Set, 5-141, 5-146
C
Certainty Equivalent, 1-4, 5-175, 5-176, 5-177
Ceteris Paribus, 1-5
Coase, Ronald, 6-220, 6-221
Cobb-Douglas Production Function, 4-82, 4-84, 4-
85, 4-90, 4-91, 4-92, 4-93, 4-96, 4-104, 5-148,
5-149, 5-153, 5-156, 5-158, 5-162, 5-184, 5-185,
5-188, 5-189, 5-191, 5-194
Compact Disc, 5-143, 7-309
Comparative Advantage, 2-36, 2-37, 2-38, 2-39, 2-
40
Comparative Statics, 1-7, 2-30, 4-87, 4-92
Compensating Differential, 5-164, 5-165
McAfee: Introduction to Economic Analysis, http://www.introecon.com, July 24, 2006
8.317
Competition
Imperfect, 7-270, 7-272, 7-273, 7-274, 7-313
Competitive Equilibrium, 6-217
Complement, 2-12, 2-15, 2-16, 2-17, 2-24, 2-25, 2-
30, 4-83, 4-89, 4-90, 4-91, 5-149, 5-156, 5-162,
5-163, 5-164, 7-292
Concave, 2-33, 5-158, 5-159, 5-160, 5-170, 5-174,
5-175, 5-181, 5-185, 7-289,7-290
Consols, 4-115
Consumer Surplus, 2-8, 2-10, 2-11, 2-13, 2-14, 2-19,
2-21, 2-30, 6-240, 6-241
Consumption, 2-8, 2-9, 2-10, 2-12, 3-56, 3-58, 3-59,
3-78, 4-126, 4-137, 5-139, 5-140, 5-143, 5-145,
5-150, 5-151, 5-152, 5-154, 5-155, 5-156, 5-160,
5-162, 5-163, 5-169, 5-170, 5-171, 5-172, 5-182,
5-189, 6-245
Contract Curve, 5-183, 5-184, 5-185, 5-186, 5-188
Contracts, 2-15, 2-38, 6-241, 7-286, 7-287, 7-306,
7-307
Convexity, 7-289, 7-290,7-291
Cooperation, 7-269, 7-270, 7-287, 7-310, 7-311
Cartel, 2-26, 7-304, 7-306, 7-310, 7-311
Tacit Collusion, 7-310
Coordination, 4-81, 7-260,7-263,7-270
Corporate Finance, 4-121
Cost, 1-2, 1-3, 1-4, 1-5, 1-6, 1-7, 2-8, 2-9, 2-14, 2-15,
2-16, 2-17, 2-18, 2-22, 2-23, 2-26, 2-34, 2-35,
2-36, 2-37, 2-38, 3-53, 3-54, 3-56, 3-61, 3-63,
3-70, 4-80, 4-84, 4-85, 4-86, 4-93, 4-94,4-95,
4-96, 4-97, 4-98, 4-99, 4-100, 4-101, 4-102, 4-
103, 4-104, 4-105, 4-107, 4-108, 4-109, 4-112,
4-114, 4-118, 4-120, 4-121, 4-123, 4-124, 4-126,
4-130, 4-131, 4-136,5-143,5-151, 5-165, 5-167,
5-172, 5-175, 5-178, 5-179, 5-180, 5-181, 6-199,
6-200, 6-201, 6-202, 6-204, 6-205, 6-206, 6-
207, 6-208, 6-211, 6-214, 6-215, 6-216, 6-217,
6-218, 6-220, 6-225, 6-226, 6-233, 6-234, 6-
235, 6-237, 6-238, 6-239, 6-241, 6-242, 6-243,
6-244, 6-246, 6-247, 6-248, 6-249, 6-250, 7-271,
7-272, 7-273, 7-277, 7-278, 7-279, 7-282, 7-283,
7-284, 7-286,7-287, 7-288, 7-289, 7-292, 7-
293, 7-294, 7-303, 7-304, 7-308, 7-309
Fixed, 4-96, 4-112, 7-273, 7-274, 7-282
Long-Run, 4-96, 6-209
Marginal, 1-7, 2-14, 2-15, 2-16, 2-18, 2-23, 2-24,
2-37, 2-38, 4-95, 4-96, 4-97, 4-98, 4-99, 4-
100, 4-104, 6-198, 6-201, 6-205, 6-213, 6-
214, 6-226, 6-232, 6-234, 6-235, 6-236, 6-
237, 6-240, 6-241, 6-242, 6-243, 6-244, 7-
271, 7-272, 7-273, 7-275, 7-277, 7-281, 7-282
Opportunity, 1-3, 1-4, 2-33, 2-34, 2-36, 2-37, 6-
198
Short-Run, 4-95, 4-96
Variable, 4-96, 4-97, 4-98, 4-99, 4-104, 4-107
Cost-Benefit Analysis, 1-2
Cournot, 7-270, 7-272, 7-273, 7-274, 7-313
D
Damages, 7-311
Dead Weight Loss, 6-198, 6-199, 6-202, 6-203, 6-
204, 6-205, 6-206, 6-211, 6-215, 6-235, 7-272
Demand
Compensated, 5-152
Constant Elasticity, 2-29, 2-30, 4-130, 6-235, 6-
237
Elasticity, 2-27, 2-29, 2-30, 4-127, 4-135, 6-200,
6-208, 6-218, 6-224, 6-225, 6-235, 6-236,
6-237, 6-238, 6-239, 6-240, 7-272, 7-273
Market, 2-18, 2-19
Perfect Substitute, 4-83, 5-149
Depreciation, 4-131, 4-135, 4-136, 4-137, 4-138, 5-
167
Differentiation, 7-279
Horizontal, 7-279
Vertical, 7-279
Diminishing Marginal Returns, 2-33, 2-37
Diminishing Marginal Value, 2-9
Discounting, 4-123, 4-127, 4-128, 4-129, 4-131, 4-
133, 4-136, 4-138, 7-268, 7-269
Disney, 6-232, 7-279, 7-312
Distribution, 3-50, 4-122, 5-178, 5-181, 6-230, 6-
245, 6-246, 7-275, 7-276, 7-297, 7-298, 7-299,
7-300, 7-301, 7-302
Dominant strategy, 7-252, 7-253, 7-254, 7-265, 7-
268, 7-296, 7-299, 7-300
DRAM, 2-8, 2-23, 4-112, 4-113
DuPont, 4-81
Durable Good, 3-78, 4-130
Dynamic Optimization, 5-169, 5-172, 5-173
E
eBay, 1-1, 2-18, 6-195, 7-295, 7-299
Economy of Scale, 4-100, 4-101, 4-102, 4-103, 4-
104, 4-109, 4-110, 4-112, 4-113, 4-114, 6-233, 7-
273
Economy of Scope, 4-101
Edgeworth Box, 5-181, 5-182, 5-185, 5-188
Education, 1-3, 2-28, 3-48, 4-120, 6-213, 6-214, 6-
218, 6-249, 6-250
Efficiency, 2-22, 2-26, 2-37, 4-93, 4-109, 4-130, 5-
182, 5-183, 5-185, 5-188, 6-195, 6-203, 6-212,
6-213, 6-216, 6-217, 6-218, 6-219, 6-220, 6-
221, 6-224, 6-225, 6-226, 6-227, 6-229, 6-230,
6-231, 6-233, 6-242, 6-246, 6-247, 6-248, 7-
273, 7-282
Elasticity, 2-29, 4-109, 4-110, 4-133, 4-135, 6-200,
6-203, 6-207, 6-208, 6-223, 6-224, 6-235, 6-
236, 6-237, 6-238, 6-239, 6-240, 7-272
Demand, 2-27, 2-29, 2-30, 4-127, 6-224, 6-235,
6-236, 7-272, 7-273
Supply, 2-30
Unitary, 2-29
Engel Curve, 5-155, 5-156, 5-157
McAfee: Introduction to Economic Analysis, http://www.introecon.com, July 24, 2006
8.318
Entrepreneur, 2-17, 2-18, 4-85, 4-86, 4-87, 4-88,
4-89,4-91,4-92,4-93,4-97
Entry, 3-64, 4-102, 6-211, 6-241, 7-251, 7-252, 7-
253, 7-273, 7-274, 7-282, 7-309
Equilibrium, 2-8, 2-20, 2-21, 2-22, 2-30, 4-93, 4-
104,4-105,4-107,4-108,4-109,4-110,4-112,
4-114, 4-133, 5-167, 5-168, 5-169, 5-187, 5-188,
5-189, 5-190, 5-191, 5-192, 5-193, 5-194, 6-197,
6-199, 6-203, 6-205, 6-206, 6-207, 6-211, 6-
216, 6-224, 6-225, 6-227, 6-230, 7-253, 7-255,
7-256, 7-257, 7-258, 7-260, 7-261, 7-262, 7-
263, 7-264, 7-265,7-268,7-269,7-270, 7-273,
7-274, 7-275, 7-276, 7-278, 7-280, 7-281, 7-282,
7-297, 7-298, 7-311
Exclusive Dealing, 7-308
Expenditure Shares, 5-149
Externality, 4-112, 6-213, 6-214, 6-215, 6-216, 6-
217, 6-218, 6-219, 6-220, 6-221, 6-224, 6-225,
6-226, 6-233, 7-278
Common Resource, 6-216
Commons, 4-127, 6-216, 6-226, 7-262
Network, 4-112, 6-233
Extinction, 6-216, 6-223, 6-224, 6-225, 6-226
Exxon-Mobil, 4-81, 7-311, 7-312
F
Factor of Production, 2-40
Factor Price Equalization, 2-40
Factors of Production, 2-38, 2-39
FedEx, 6-232
Firm, 4-79, 4-97, 7-265, 7-268, 7-269, 7-270, 7-
292
Competitive, 4-97, 4-98, 4-104
Corporation, 1-5, 4-79, 4-80, 4-81, 7-283, 7-306
Non-Profit, 4-80
Partnership, 4-79, 4-80
Proprietorship, 4-79, 4-80
Fixed-Proportions, 4-83
Fluctuations, 3-77, 3-78, 7-288
Ford, 1-1
Foreclosure, 1-2, 7-312
Free Market, 6-195, 6-207
Free-Rider, 6-228, 6-229
G
Gains from Trade, 2-10, 2-22, 2-37, 2-39, 4-131, 6-
198, 6-201, 6-202, 6-203, 6-205, 6-206, 6-221,
6-240, 6-241, 6-242, 6-246, 6-247, 6-248, 7-
272, 7-286
Game Theory, 7-251, 7-256
Battle of the Sexes, 7-256, 7-259, 7-267, 7-270
Common Knowledge, 7-270
Elimination of Dominated Strategies, 7-253, 7-
254, 7-255, 7-257
Folk Theorem, 7-269, 7-270
Grim Trigger Strategy, 7-268, 7-269, 7-270, 7-
310
Mixed Strategy, 7-258, 7-259, 7-260, 7-261, 7-
263, 7-264, 7-265, 7-275
Pure Strategy, 7-258, 7-259, 7-260, 7-261, 7-
262,7-263, 7-264, 7-265,7-275
Second-Mover Advantage, 7-267
Strategic Behavior, 7-251
Subgame Perfection, 7-266, 7-267, 7-268, 7-269,
7-270, 7-284
Gasoline, 1-2, 2-8, 2-11, 2-16, 2-24, 2-26, 4-81, 4-
108,4-109, 5-178, 6-195, 6-218,7-303
General Electric, 7-309, 7-311
General Equilibrium, 5-188, 5-189
Welfare Theorems, 5-188
General Motors, 4-81
Chevrolet, 2-28
Goldwyn, Samuel, 7-301
Government, 1-1, 1-2, 1-5, 3-48, 3-56, 3-58, 3-59,
3-65, 3-66, 3-67, 3-68, 3-69, 3-70, 3-71, 3-72,
4-80, 4-114, 4-115, 5-160, 5-171, 5-177, 6-195,
6-197, 6-198, 6-199, 6-200, 6-202, 6-210, 6-
211, 6-213, 6-218, 6-219, 6-220, 6-232, 6-237,
7-295, 7-296, 7-300, 7-303, 7-305, 7-311, 7-312,
7-314
Greyhound, 7-314
Gross Domestic Product, 3-56, 3-58, 3-59, 3-61, 3-
65, 3-66, 3-67, 3-70, 3-71, 3-72, 3-73, 3-74, 3-
75
H
Holmstrom, Bengt, 7-292
Homo Economicus, 1-5
Homogeneity, 6-231, 7-289, 7-290, 7-291, 7-292
Homogeneous Function, 4-104, 7-289, 7-291, 7-294
Hotelling Model, 7-279, 7-280, 7-282
Hotelling, Harold, 7-279
Hypothesis, 5-152, 6-221, 7-257, 7-297
Hysteresis, 2-26
I
IBM, 4-101
Identification, 6-238
Income, 1-3, 2-11, 2-13, 3-50, 3-51, 3-52, 3-53, 3-
54, 3-55, 3-56, 3-58, 3-59, 3-64, 3-71, 3-72, 3-
73, 4-105, 5-139, 5-143, 5-150, 5-152, 5-154, 5-
155, 5-156, 5-158, 5-160, 5-161, 5-163, 5-164, 5-
169, 5-171, 5-172, 5-173, 6-195, 6-203, 6-238
Income Effect, 5-152, 5-156, 5-158, 5-163, 5-171
Indifference Curve, 5-143, 5-144, 5-145, 5-146, 5-
147
Inferior Good, 2-11, 5-155, 5-156, 5-157
Information, 2-23, 3-56, 3-57, 4-101, 4-120, 6-245,
6-246, 7-251, 7-278, 7-295, 7-298, 7-299, 7-
300, 7-301, 7-302, 7-303, 7-304, 7-305, 7-311
McAfee: Introduction to Economic Analysis, http://www.introecon.com, July 24, 2006
8.319
Intertemporal Budget Constraint, 5-169, 5-172, 5-
173
Inventory, 2-21, 4-136, 4-137
Isocost, 4-94
Isoquant, 4-82, 4-94, 5-149, 5-152, 5-182, 5-183,
5-185, 5-187, 5-188
M
Managers, 4-80, 4-101, 7-310
Marginal Product, 4-84, 4-89, 4-103, 4-104
Value of, 4-84, 4-85, 4-86, 4-92, 4-103
Marginal Rate of Substitution, 5-144, 5-146, 5-148,
5-157
Marginal Rate of Technical Substitution, 4-94
Marginal Returns, Diminishing, 2-33, 2-37
Marginal Revenue, 6-234, 6-235
Marginal Value, Diminishing, 2-9
Market for Lemons, 6-246
Marketing, 4-81
Markets, 1-1, 1-2, 1-6, 2-21, 2-24, 2-25, 4-113, 4-
126, 6-195, 6-213, 6-246, 6-248, 7-274, 7-278,
7-295, 7-305, 7-311, 7-312, 7-313, 7-314
Maximum Sustainable Yield, 4-129, 4-130
McAfee, R. Preston, i, iii, 4-81, 5-175, 6-246
Mercedes, 7-309
Merger, 7-308, 7-309, 7-311, 7-312, 7-313, 7-314
Horizontal, 7-312
Microsoft, 1-2, 2-24, 6-232, 6-233, 7-253, 7-254, 7-
307, 7-309
Milgrom, Paul, 7-292, 7-303
Monopoly, 6-195, 6-201, 6-202, 6-232, 6-233, 6-
234, 6-235, 6-236, 6-237, 6-241, 6-244, 7-251,
7-270, 7-272, 7-273, 7-274, 7-277, 7-307, 7-
308, 7-310, 7-313, 7-314
Monopoly Power, 6-232, 6-236, 7-273, 7-274, 7-
307, 7-313
Morgenstern, Oskar, 7-251
Mortgage, 4-115, 4-116, 4-124, 4-125
Multi-Tasking, 7-288, 7-291
Myerson, Roger, 6-246, 6-248
Myerson-Satterthwaite Theorem, 6-246, 6-248
N
Nash, John, 7-251, 7-255
Netscape, 1-2, 7-307
Network, 4-112, 6-233, 7-312
Newton, Sir Isaac, 6-221
Nonexcludability, 6-226
Nonrivalry, 6-226
Normal Good, 5-155, 5-156, 6-238
Northwest, 2-38, 2-40
NPV, 4-117, 4-118, 4-119, 4-120, 4-121, 4-122, 4-
123, 4-124, 4-125
Numeraire, 5-143, 5-190
0
Ohlin, Bertil, 2-40
Oil, 1-2, 2-8, 2-16, 2-17, 2-25, 2-26, 3-60, 3-72, 4-
81, 4-108, 4-118, 4-120, 4-121, 4-124, 4-125, 4-
126, 4-127, 4-135, 5-164, 6-216, 6-218, 7-295,
7-296, 7-301, 7-303, 7-304, 7-306, 7-311
Oligopoly,'7-270, 7-314
OPEC, 2-26
Opportunity Cost, 1-3, 1-4, 2-33, 2-34, 2-36, 2-37,
6-198
Option Value of Investment, 4-121
Options, 1-4, 4-121, 4-122, 4-124, 4-125
O'Rourke, P. J., 7-283
P
Pareto Efficiency, 5-182, 5-183, 5-184, 5-188
Partnership, 4-79, 4-80
Patent, 4-85, 6-232, 6-246
Perron-Frobenius Theorem, 5-190, 5-191
Peter Principle, 4-101
Pigou, Arthur Cecil, 6-217
Porsche, 2-11, 4-130, 4-131, 6-232
Post Office, US, 6-232
Preferences, 1-2, 2-13, 4-125, 5-139, 5-140, 5-145,
5-147, 5-148, 5-165, 5-167, 5-168, 5-169, 5-177,
5-182, 5-185, 5-193, 5-194, 6-213, 6-229, 6-
230, 6-231, 6-245, 7-264, 7-270, 7-279
Present Value, 4-114, 4-115, 4-116, 4-117, 4-118, 4-
119, 4-120, 4-126, 4-127, 4-133, 4-136, 5-167, 5-
169, 5-170, 5-172, 6-249, 7-269, 7-310
Price
Ceiling, 6-203, 6-205, 6-206, 6-207
Hedonic, 1-4
Law of One, 7-274, 7-278
Peak Load, 6-243, 6-244
Reservation, 2-13, 5-178, 5-179, 5-180, 5-181
Price - Cost Margin, 7-272, 7-273, 7-313
Price Discrimination, 6-237, 6-238, 6-239, 6-240,
7-275, 7-308, 7-314
Coupons, 6-239
Direct, 6-239
Indirect, 6-239
Quantity Discount, 6-239
Price Dispersion, 7-275
Price Floor, 6-202, 6-203, 6-205, 6-206, 6-207, 6-
210, 6-211
Price Support, 6-210, 6-211
Price System, 2-37, 2-38, 5-186, 5-187, 5-188, 6-
238, 6-245
Price-Cost Margin, 6-235, 7-271
Pricing
Ramsey, 4-125, 4-126, 6-244
Private Values, 7-295, 7-296, 7-297, 7-298, 7-299,
7-300, 7-301, 7-304
Production Possibilities Frontier, 2-32, 2-33, 2-34,
2-35, 2-36, 2-37, 2-38
Property Rights, 6-220, 6-221
McAfee: Introduction to Economic Analysis, http://www.introecon.com, July 24, 2006
8.320
Proprietorship, 4-79, 4-80
Public Good, 6-226, 6-228, 6-229, 6-230, 6-231, 7-
262
Q
Quintiles, 3-50, 3-51
R
Raising Rivals' Cost, 7-312
Ramsey Pricing, 4-125, 6-244
Ramsey, Frank, 4-125
RE/MAX, 7-287
Reagan, Ronald, 3-65
Regulation
Price-Cap, 6-242
Rent Control, 6-207, 6-208, 6-209
Rents, 4-85, 6-208
Research and Development, 3-72, 4-101, 4-102, 7-
282
Reservation Price, 2-13, 5-178, 5-179, 5-180, 5-181
Revealed Preference, 4-87, 4-88, 4-89, 4-90, 6-237
Revenue equivalence, 7-298
Risk, 1-4, 2-8, 4-85, 4-101, 4-114, 4-118, 4-119, 4-
120, 4-121, 4-125, 4-138, 5-174, 5-175, 5-176, 5-
177, 5-181, 6-211, 7-283, 7-284, 7-286, 7-287, 7-
288, 7-299,7-303, 7-304, 7-305, 7-310
Certainty Equivalent, 1-4, 5-175, 5-176, 5-177
Risk Management, 1-4, 4-101
Risk Aversion, 5-176, 5-177, 7-287, 7-299, 7-304
Risk Premium, 1-4, 5-175, 5-176, 5-177, 5-178
Rivalry, 4-81
Roosevelt, Teddy, 7-308
S
S&P 500, 5-177
Samuelson, Paul, 2-40
Satiation, 5-150
Satterthwaite, Mark, 6-246
Saving, 2-26, 4-115, 4-127
Scale Economy, 4-102, 6-233
Schelling, Thomas, 7-251
Schwarzenegger, Arnold, 5-139
Sears, 4-81
Second-Price Auction, 7-299
Self-Interest, 1-5, 1-6
Selten, Reinhart, 7-266
Service, 2-20, 2-37, 3-63, 3-68, 4-85, 5-144, 6-
202, 6-203, 6-231, 6-233, 6-242, 7-253, 7-308,
7-312
Shadow Value, 4-91, 4-92
Sharecropper, 7-287
Shortage, 2-8, 2-20, 2-21, 4-126, 5-164, 6-206, 6-
207, 6-208, 6-209
Signaling, 6-248, 6-249, 6-250
Smith, Adam, 2-35, 2-36
Sony, 1-1, 5-178, 6-233
Spence, Michael, 6-248
Standard Oil, 4-81, 7-306, 7-309
Standardization, 4-79, 6-233, 6-250, 7-274
Statistics, 3-41, 3-57, 3-76, 6-250
Correlation, 5-193, 5-194
Normal Distribution, 5-177, 7-301, 7-302
Steady State, 2-20, 6-222, 6-223, 6-224, 6-225, 7-
256
Strategy, 4-97, 4-121, 4-122, 4-123, 4-133, 5-160,
6-206, 6-218, 6-220, 6-226, 6-242, 7-251, 7-
252, 7-253, 7-254, 7-255, 7-256, 7-257, 7-258,
7-259, 7-260, 7-261,7-263,7-268, 7-269, 7-
270, 7-296, 7-298, 7-299, 7-310
Substitute, 2-11, 2-12, 2-17, 2-24, 2-25, 4-89, 4-
110, 6-203, 6-208, 6-238, 7-272, 7-294, 7-307,
7-313
Supergames, 7-268, 7-269
Supply, 1-6, 2-8, 2-13, 2-14, 2-15, 2-16, 2-17, 2-18,
2-19, 2-20, 2-21, 2-22, 2-23, 2-24, 2-25, 2-26,
2-30, 2-31, 2-32, 2-37, 3-61, 4-103, 4-104, 4-
105, 4-107, 4-108, 4-109, 4-110, 4-112, 4-113, 4-
114, 4-126, 4-129, 4-130, 4-131, 4-137, 5-139, 5-
143, 5-160, 5-162, 5-163, 5-165, 5-170, 5-187, 5-
188, 5-189, 5-190, 5-193, 6-195, 6-196, 6-197, 6-
198, 6-199, 6-200, 6-201, 6-202, 6-203, 6-205,
6-207, 6-208, 6-209, 6-211, 6-216, 6-217, 6-
219, 6-235, 6-237, 7-292, 7-304
Complement, 2-16, 2-17, 2-25
Constant Elasticity, 2-30
Elasticity, 2-30, 6-200
Increase in, 2-16, 2-23, 2-25
Market, 2-19, 4-104
Substitute, 2-17, 2-25
Surplus, 2-8, 2-10, 2-13, 2-20, 2-21, 3-73, 5-164, 6-
198, 6-203, 6-210, 6-211, 6-241
T
Tax
Ad Valorem, 6-195, 6-201
Excess Burden of, 6-200, 6-202
Excise, 3-72, 3-73, 6-195
Income, 3-72, 3-73, 6-195
Sales, 1-3, 3-73, 4-80, 6-195
Time preference, 5-169, 5-170
Trailways, 7-314
Transaction Costs, 4-114, 6-195
Truman, Harry, 1-5
Tyranny of the Majority, 6-209
U
U. S. Federal Trade Commission, 7-306, 7-308, 7-
309, 7-311, 7-312
U.S. Department of Justice, 1-2, 7-306, 7-307, 7-
309, 7-310, 7-311, 7-312
Uncertainty, 2-26, 4-120, 4-121, 4-136, 7-305
Unitization, 6-216
McAfee: Introduction to Economic Analysis, http://www.introecon.com, July 24, 2006
8.321
Utility, 5-139, 5-140, 5-143, 5-144, 5-146, 5-147, 5-
148, 5-149, 5-150, 5-151, 5-152, 5-153, 5-154, 5-
158, 5-159, 5-160, 5-162, 5-167, 5-168, 5-170, 5-
174, 5-175, 5-176, 5-181, 5-182, 5-183, 5-184, 5-
185, 5-186, 5-187, 5-188, 5-189, 5-191, 5-194, 6-
228, 6-229, 6-247, 7-284, 7-286, 7-287, 7-290,
7-293
Compensated Demand, 5-152
Convex Preferences, 5-145, 5-147, 5-148
Indifference Curve, 5-143, 5-144, 5-145, 5-146, 5-
147
Quasilinear, 5-159
Satiation, 5-150
von Neumann-Morgenstern, 5-174
V
Value
Marginal, 1-7, 2-9, 2-10, 2-9, 2-11, 2-14, 2-15, 4-
89,4-131,4-132,4-133, 5-162, 5-167, 5-170,
6-196, 6-201, 6-214, 6-218, 6-242
Vickrey, William, 7-299
Video Home System, 2-23, 6-233
Vincent, Daniel, 5-175
von Neumann, John, 7-251
Voting
Median Voter, 6-230
w
Warranties, 6-246
Welfare Analysis, 1-2
Western Electric, 4-134, 4-135
Willingness To Pay, 1-2, 2-8, 2-11, 2-13, 6-197, 6-
238, 6-239, 7-295
McAfee: Introduction to Economic Analysis, http://www.introecon.com, July 24, 2006
8.322