75

CHAPTER 11

WHY NOT SPLIT THE CHECK?

A Briefing on Market Failure

Last Friday, somewhere near you, a group of students slurped root beers at a popular burger joint. Frosted mugs cluttered the table like miniature Manhattan skyscrapers. Gossip and wisecracks filled the air. When the check arrived, a student in blue jeans and a black concert t-shirt called out, “Let’s just split it.” Those with the most mugs before them greeted the motion with nods. Those with fewer mugs had blank faces. One light drinker meekly mumbled something about fairness. An equal splitting of the check seemed inevitable. Then a baseball player sitting in the corner leaned up to the table, his face illuminated by the restaurant’s blinking red neon light. “My friends,” he inserted into the awkward silence, “to split the check isn’t only an injustice to those who had fewer drinks; it also creates a strong incentive for overindulgence.” The others were perplexed. But as he leaned back, the light illuminated his economics society t-shirt and they knew that logic was on his side. This chapter explains the error of these students’ ways and reveals other sources of market failure that surround us daily and confound many of our stabs at efficiency.

Ideal market conditions yield efficient allocations of inputs and outputs. Market failure is the reason why some markets do not bring about the best outcomes for society. As explained in Section 11 Economics by Example, perfectly competitive firms achieve market efficiency by equating marginal costs and marginal benefits. This chapter provides a constructive overview of the sources of market failure: externalities, imperfect competition, imperfect information, and public goods. Each of these sources receives further mention later in the book: Imperfect information is the focus of Chapter 12; imperfect competition is central to the issues in Chapters 13 and 14; public goods are the topic of Chapter 18; and externalities are reprised in Chapter 29. As you’re about to learn, some of the solutions to market failure are as simple as paying for your own root beers.

76

Externalities are costs or benefits felt beyond, or external to, the people causing them. When you get a flu shot, you create positive externalities because other people won’t catch the flu from you. When you drive a car, you create the negative externalities of road congestion and air pollution. To emit toxins into the air or water during a manufacturing process, smoke in public, play loud music, catch fish that other fishers are angling for, or paint your house a dreadful shade of green is to impose external costs—negative externalities—on other people.

You may not think of root beer as a potential source of negative externalities, but when you buy a root beer with a group that is splitting the check equally, you impose a cost on everyone in that group.¹ Suppose a root beer costs $2 and you have 10 people in your group. Suppose also that diminishing marginal utility leads you to value your first 5 root beers at $3, $2.50, $1.25, $0.50, and $0.10, respectively. If you were not splitting the check, you would buy 2 root beers for $2 each—the first that’s worth $3 to you and the second that’s worth $2.50. You wouldn’t pay $2 for a third root beer that’s worth $1.25. The situation changes when you’re splitting the check 10 ways. In that case each root beer raises the tab by $2, but it raises your share of the tab by only $2/10 = 20 cents. Thus, it’s rational for you to purchase 4 root beers—all of those that are worth at least 20 cents to you—even though the last 2 are worth less than the $2 price that the group pays.

¹ For more on the negative externality problem with splitting the check, and for empirical evidence of the problem, see www.gsb.uchicago.edu/fac/uri.gneezy/vita/Restaurant.pdf.

The problem grows because everyone in the group faces the same incentive to overconsume. If your friends share your preference for root beer, each of the 10 friends will consume 4 beverages, bringing the total bill up to $80. Each friend’s share of the bill will be $8—twice what they would have paid on their own and 75 cents more than the $3 + $2.50 + $1.25 + $0.50 = $7.25 value of the root beers to them. Notice that you cannot improve on this situation by reducing your order. If you ordered 3 root beers, you would reduce your tab by $0.20 and you would receive $0.60 cents less worth of beverage. By splitting the check, your group creates externalities that thwart efficient decisions.

Externality problems can be resolved if ways can be found to have everyone internalize, or feel for themselves, the full costs or benefits of their decisions. In the root beer example, that means having everyone pay his or her own bill. The negative externalities from goods such as gasoline, cigarettes, and alcohol are internalized if excise taxes (taxes levied on particular goods such as these) approximate the associated external costs. A solution to the underconsumption that accompanies positive externalities is for the government to subsidize purchases, as is the case for flu shots as well as education, tree planting, and clean energy.

77

Because goods creating positive externalities are underconsumed and goods creating negative externalities are overconsumed from a societal standpoint, another solution is to require people to purchase the efficient quantity of these goods. It is common for governments to require a particular number of childhood immunizations and education through high school while limiting the volume of power plant emissions and the number of pets that can be kept in a household. For example, in the interest of neighborhood serenity and sanitation, no resident in the city of Grand Junction, Colorado, may keep more than 3 adult pets, such as dogs or cats. Chapter 26 elaborates on the role of government and explains how the enforcement of private property rights can assist with externality problems.

Hawaiian shirts sell for $30 and up in the gift shop of the Honolulu International Airport. In downtown Honolulu, the Daiei supermarket sells similar Hawaiian shirts for $14. Both places sell shirts that are made in Hawaii at comparable costs; most of the price differential stems from differing levels of competition. Daiei faces competition from hundreds of other shirt sellers, and if the managers were to raise their prices significantly above marginal cost, competitors would undercut their prices and steal away some of their customers and profits. That’s how prices in competitive environments are brought down to approximately the marginal cost of production. When price equals marginal cost, consumers keep buying until their marginal benefit equals the price, and because price equals marginal cost, the efficiency condition of marginal cost equaling marginal benefit is satisfied. As explained in further detail in Section 11 Economics by Example, competitive markets can achieve productive, allocative, and distributive efficiency.

Prices are higher at the airport because the number of sellers is limited² and their clientele is captive. Many of the airport customers are leaving the island and have no choice but to make any additional purchases at the airport gift shop. That gives the gift shop market power—the ability to influence prices—in that location. It is market power that allows a single operator of ballpark concession stands to charge relatively more for hot dogs and allows a campus bookstore to charge more for batteries than do competing stores in a distant shopping area. In the absence of competitors who would try to undercut excessive prices, the gift shop can charge prices above the marginal cost of production. Suppose for simplicity that it costs $14 to put each Hawaiian shirt on the market (that is, the $14 downtown price equals the marginal cost, as is the case in perfectly competitive markets). Different consumers place a variety of values on Hawaiian shirts, from 0 to more than $100. When the price is $30, consumers who value the shirts in the range between $14 and $30 are dissuaded from purchasing shirts, even though they value shirts at more than the $14 cost of production. For example, when someone who values a shirt at $20 does not buy 1 because it costs $30, deadweight loss occurs because society loses the $20 − $14 = $6 net benefit that the shirt could have provided. Imperfect competition leads to market failure as a result of these lost opportunities for net gains.

² It is common for airports to offer exclusive “master concession” agreements to retailers. For example, HMSHost Corp. owns master concession rights for all food and beverage operations at Honolulu International Airport. See http://starbulletin.com/2002/08/29/business/story1.html.

78

McDonald’s Restaurant and Doug’s Restaurant are both located on North Canal Boulevard in Thibodaux, Louisiana. Why would visitors to Cajun country eat at McDonald’s when they could encounter local cuisine at Doug’s? One reason involves imperfect information, which exists when at least one buyer doesn’t know the true benefit of a good or service or at least one seller doesn’t know the true cost of providing a good or service. As far as visitors know, Doug’s may or may not be a good place to eat. There is relatively little uncertainty about a meal at McDonald’s because most visitors know exactly what’s on the menu and how it will taste. The additional investment of time and money that customers must make to determine food quality at an unknown restaurant stands in the way of competition from new, no-name-brand competitors, such as Doug’s, and thereby thwarts market efficiency. More broadly, although plenty of businesses charge prices above marginal cost, the difficulty and expense of acquiring anything close to perfect information stand in the way of competition from new entries, such as Bob’s Bank, Carl’s Cars, and Dave’s Diner, which lose customers because of uncertainty about quality.

When buyers and sellers don’t share the same information, the problem of asymmetric information arises. As explained in Chapter 5, those in the marriage market use ease of attraction as a signal of the unseen qualities of a potential mate. Employers and shoppers likewise use college degrees, grades, name brands, and prices as imperfect signals of quality among workers and products when perfect information is not available. Nobel laureate George Akerlof pointed out that there is asymmetric information between buyers and sellers of used cars: Sellers know things about the quality of their cars that potential buyers do not. Suppose that used cars are either good or bad and that good used cars are worth $20,000, whereas bad used cars are worth $10,000. Given the uncertainty about whether a car will be a “lemon” (bad), buyers are unlikely to pay $20,000 for a used car. If, in the buyer’s estimation, there is a 20 percent chance that the car is a lemon and an 80 percent chance that it’s worth $20,000, the expected value of the car (the sum of each probability times the outcome that occurs with that probability) is (0.2)($10,000) + (0.8)($20,000) = $18,000. Thus, sellers of good used cars won’t get fair prices for their cars, but sellers of lemons will. The result is that owners of good used cars are more likely to hold on to their cars and the used car market may be made up primarily of lemons.

In response to the risk that the market for used cars might become a market for lemons, many buyers of used cars (and homes and boats) invest in a professional evaluation of their contemplated purchase to try to gain more of the information held by the seller. Car buyers are also wise to request a vehicle identification number (VIN) check on their car at a Web site such as www.carfax.com, which will reveal information on odometer readings, the number of previous owners, any major accidents in the vehicle’s past, and similar potential problems that sellers know about and buyers otherwise don’t.

79

Adverse selection occurs when it costs varying amounts of money to serve different customers but sellers cannot distinguish between high-need and low-need buyers. It would make sense for people who take more physical risks to buy more health insurance; for big eaters to patronize all-you-can-eat restaurants; for people with terminal diseases to buy more life insurance; and for employees with larger families to seek employers who provide better family health, vision, and dental benefits. In efforts to limit adverse selection, insurance companies typically require physical exams and may not cover preexisting illnesses. All-you-can-eat restaurants and employers have a harder time dodging the adverse-selection bullet because they cannot discriminate on the basis of appetite or family size.

Moral hazard is the tendency for those with insurance against a problem to take fewer precautions to avoid that problem. Knowing that a hospital stay can cost $2,000 or more per day, people with insurance may be more likely to take up skydiving than are people who have to shoulder their entire medical bill themselves. Laid-off workers covered by unemployment insurance take more time to find new jobs than do those who are ineligible or whose coverage has lapsed. And people with fire insurance may take more risks with candles in their homes because they will not have to pay the entire price of a new house if there is a fire. Insurance companies would relish the ability to better monitor the risk-taking behavior of their customers. A more practical but imperfect solution is for insurers to increase copayments and experienced-based premiums to boost the share of the burden that insured people must pay when their precautions turn out to be inadequate. People respond to incentives in rational ways, so the more they have to pay for their mistakes, the less likely they will be to make mistakes.

Clearly, the solutions to imperfect information are themselves imperfect and costly. Car and human checkups intended to level the informational playing field sometimes do not. High copayments place a financial burden on the poor. Only a 100 percent copayment would bring decision makers to internalize the full costs of their behaviors, and that would eliminate the risk-sharing benefits of insurance. When solutions are unavailable or incomplete, the information problems explained in this section obscure true marginal benefits or marginal costs and deter the efficient convergence of the two.

A handful of spy satellites spin 100 miles above the earth, recording a constant video stream of activities below. These billion-dollar detectives are intended to make the world safer by tracking military, drug, and terrorist activities. The U.S. National Reconnaissance Office operates satellites to intercept messages, draw maps, and gather visual information for U.S. intelligence agencies.³ United Nations inspectors used satellites to get an advance look at possible nuclear facilities in Iraq. And satellites help the U.S. Drug Enforcement Agency peek into the hideouts of druglords in Cali, Colombia, before a raid. As useful as spy satellites are to society, it would be difficult for a spy satellite market to bring about the efficient number of them because there are incentives for beneficiaries to receive the benefits of satellites without helping out with the costs.

³ See www.spacetoday.org/Satellites/YugoWarSats.html.

80

A few relevant terms are useful in explaining why this is the case. A good or service is nonrival in consumption if 1 person’s consumption of it does not affect anyone else’s consumption of it. A street band is nonrival in consumption because 1 passerby’s enjoyment of the music doesn’t prevent others from enjoying it. In contrast, shoes are rival in consumption because if 1 person is wearing them another person cannot. A good or service is nonexcludable if it is impossible to prevent other people from benefiting as long as 1 person is benefiting from it. A streetlight is nonexcludable because, if it is lighting the way for 1 driver, other drivers can benefit from the same light. Dorm rooms are excludable because you can lock your door and prevent others from entering. Goods and services that are both nonrival in consumption and nonexcludable are called public goods.

The challenge of creating the efficient number of spy satellites, as with street bands and streetlights, is that they are public goods. Everyone can benefit from them no matter who pays for them. Thus, the temptation is to be a free rider and enjoy the benefits of someone else’s expenditures. Some street bands appear because some people give them tips, but more would appear if everyone who enjoyed the music chipped in. If the Central Intelligence Agency (CIA) went from door to door asking everyone to pitch in to pay for the next satellite, it would be rational from an individual standpoint for many people to contribute less than their fair share of the price in hopes that others would purchase the satellites to the benefit of all. The result is that private markets for public goods generally don’t bring about the efficient quantity of the goods in question.

A common solution to the free-rider problem is for the government to collect taxes from all the beneficiaries of public goods and use the revenues to fund the appropriate quantities of them. The spy satellites are indeed paid for with tax dollars, as are streetlights, police and fire departments, the Environmental Protection Agency, and the military, among other public goods. Given the unequal use of public goods by taxpayers and the difficulty of estimating the ideal quantity of public goods, private solutions can be appealing. In some cases technological advances help to turn public goods into private goods with no free-rider problem. For instance, uncongested roads are public goods (congested roads become rival in consumption). Devices called Smart Tags can be attached to car windshields or license plates to communicate electronically with computers at toll booths and deduct tolls from the accounts of drivers, who need not stop as they pass through a toll lane. For drivers on the Dulles Greenway in Virginia, for example, this system takes much of the hassle out of toll booths and increases the viability of toll roads that require everyone to pay their fair share of highway costs.

Entire books have been written about each of these types of market failure. In terms of the primary example in this chapter, it is interesting to note that the local root beer stand isn’t the only place where the check is split. In 2005, health-care costs amounted to 15 percent of gross domestic product, up from 5 percent in 1960. Although expensive high-tech advances, doctors, malpractice lawyers, and insurance companies may all share the blame for higher prices, remember that hospitals are another place where people typically split the check. The great benefit from insurance is that it allows them to spread risks and rising costs. The downside is that, as with root beer, people make decisions about additional health-care purchases knowing that they will personally bear only a small fraction of the price of these services. Since 1960, the share of health-care costs paid out of pocket dropped from 55 to 16 percent.⁴ According to John C. Goodman, president of the National Center for Policy Analysis, “the primary reason health care costs are rising is that most spending on health care is done with someone else’s money rather than the patient’s.”⁵ If the doctor asked you whether you’d like to use an X-ray to rule out the possibility that the frozen pizza you dropped on your foot broke a toe, you might consider your copayment for the procedure rather than the full cost of an X-ray and request X-rays that are not justified by their true cost. Similarly, if your car is dented by an errant baseball, you might weigh the benefit of getting it fixed against the amount by which your premiums will go up rather than the true cost of the repair. As with the moral hazard problem of insufficient care taken by those with insurance, the problem of splitting the insurance check could be solved by charging consumers higher copayments so that they’d bear more of the cost of their decisions.

⁴ See http://aspe.hhs.gov/health/MedicalExpenditures/index.shtml.

⁵ See www.econlib.org/library/Enc/HealthInsurance.html.

81

This chapter highlights situations in which market incentives and the decisions they drive run afoul of efficiency. When individuals don’t bear the cost of their personal consumption choices, the incentive is to buy too much of goods and services that place burdens on others and too little of goods and services that benefit others. When barriers to competition exist, people can expect prices above marginal cost and lost opportunities to produce goods that are valued beyond their production cost. When information is imperfect, people cannot respond appropriately to true marginal costs and marginal benefits. And when goods are nonrival in consumption and nonexcludable, people have incentives to minimize expenditures on these public goods in hopes of free riding on the purchases of others. These problems prevent consumption from continuing until the benefit from another unit no longer exceeds the cost; as a result, the market fails to create all the net benefits that it otherwise could.