The Great Reversal

by Thomas Philippon

  More broadly, we can argue that competition increases economic freedom. In a competitive labor market, workers have the freedom to quit and find a better job. When employers compete, they offer more options to workers: different jobs, different hours, and different benefits. Labor market competition is the best defense against employers abusing and bullying their employees.

  Should we worry about the loss of competition? After all, if competition is so wonderful, why would it be threatened? Shouldn’t we all agree that competition is beneficial and team up to defend it?

  Writing more than fifty years ago, Mancur Olson (1971) explained why a spontaneous defense of competition is unlikely. Before Olson, the common wisdom was that if we all shared a common interest, we would act collectively to achieve it. Unfortunately, there is a flaw in that argument. To understand the logic of collective action, we first need to ask who gains and who loses. In the case of competition, it is rather clear. Competition destroys rents and is therefore the enemy of rent seekers.

  The key point about rents is that they are usually concentrated. When a dominant firm lobbies to prevent entry by competitors, it is protecting its own rents and perhaps the rents of a couple other dominant firms. The winners are few, and they have a lot to protect. On the other hand, the set of economic actors who are likely to get hurt if lobbying succeeds in limiting competition is widely dispersed. The consumers who buy from the industry are directly affected. Moreover, since high prices lower consumers’ real disposable income, they are less able to spend on other goods and services. Most other industries are therefore indirectly affected. These costs, however, are hidden and diffuse. Consumers may never know that entry was restricted, and the indirect costs are small, so it does not pay for anyone to pick up the fight. As a result, the households and other businesses that are hurt by rent seekers are unlikely to create their own lobby to fight back. They would certainly be happy if someone else made the effort to lobby in favor of free entry. But as individuals, they have little incentive to do it themselves. This is Olson’s central argument: concentrated special interests are likely to organize and fight to protect their rents, while diffuse majority interests are trumped. The essence of the problem is free-riding and the fact that free-riding incentives grow with the size of the group.

  The astute reader might recognize that the same argument applies to free trade among nations. International trade also creates diffuse winners and concentrated losers. Job losses and factory closings are salient facts that are likely to be picked up in the media. Lower prices are much harder to identify and job creation from trade is spread over many locations and industries. This is why free trade needs to be protected.

  But, as I have argued above, the argument for domestic competition is stronger than the argument for free trade because the benefits from free trade are more difficult to share among citizens. When trade shifts production overseas, these jobs are literally gone. The same is not true with domestic competition.

  To conclude, I would argue that the case for free competition within a country is as strong as any case one can make in economics. Unfortunately, the virtue of competition—that its positive effects are widespread—is also its downfall: the winners are dispersed, and the losers are concentrated. This is why we see a lot of lobbying aimed at restricting competition and little advocacy to protect it. We will delve deeper into this topic in the second half of the book.

  But for now, let’s figure out how to measure competition.

  * * *

  a  Bundling and competition can have surprising effects. Gregory S. Crawford, Oleksandr Shcherbakov, and Matthew Shum (2018) find an overprovision of “quality” in cable television markets. They argue that it results from the presence of competition from high-end satellite TV providers: without the competitive pressure from satellite companies, cable TV monopolists would instead engage in quality degradation. Quality overprovision implies cable customers would prefer smaller, lower-quality cable bundles at a lower price.

  b  I write “growth” because this is what most people have in mind when they think about a country becoming richer, but this needs some clarification. In the standard economic model competition has an impact on the level of GDP but not on its long-term rate of growth. Imagine a change in policy that leads to an increase in competition in domestic markets. This leads to a temporary increase in the rate of growth of the economy. Afterward, GDP remains permanently higher than it would have been without the policy change, but it eventually grows at the same rate as before because the long-term rate of growth of income per capita depends only on technological progress. Competition can have a permanent impact on growth if it encourages technological innovation. This is a hotly debated issue. The evidence suggests that competition induces more innovation, but there is no consensus on the size of this effect. We will discuss the links between competition, investment, and productivity in Chapter 4.

  CHAPTER 2

  Bad Concentration, Good Concentration

  TO ASSESS THE DEGREE of competition in an industry, economists look at three main variables: the degree of concentration (that is, whether there are lots of small firms or whether the industry is dominated by a few large firms); the profits that these firms are making, and the prices that customers pay. Each one of these variables contains useful information, but none is a perfect indicator. Taken together, however, they can give us a fairly clear picture of what is happening.

  Concentration is a bit like cholesterol; there is a good kind and a bad kind. The bad kind occurs when incumbents in an industry are allowed to block the entry of competitors, to collude, or to merge for the primary purpose of increasing their power over market-wide pricing. The good kind is when an industry leader becomes more efficient and increases its market share. In economics, concentration is often a bad sign, but not always. As an indicator of competition, it should always be taken with a grain of salt. And it should always be considered together with profits and prices.

  We are going to start by discussing the concept of market power. We will then review a few examples of deregulation where all the indicators move in the same direction. We will then study cases that are more difficult to interpret and discuss the expansion of Walmart and Amazon.

  Market power plays a central role in this book. My central argument is that there has been a broad increase in market power across the US economy, and that this increase has hurt US consumers. We therefore need to understand the causes and consequences of market power. To do so we are going to look at a couple of examples representing stylized markets.

  Market Power Versus Demand Elasticity

  Market power is a key concept in economics. It measures the ability of a firm to raise its price and increase its profits at the expense of its customers. Clearly, that can happen only if the customers in question do not have readily available alternatives. If they did, they would react to any price increase by switching to another producer. In economics, we say that market power depends on the elasticity of demand.

  To understand the relationship between market power and the elasticity of demand, consider the following example. Suppose that you want to fly from A to B. Suppose the average cost of flying one person from A to B is $200. This cost covers the wages of pilots, fuel, take-off and landing rights, airport and screening fees, and the wear and tear on the airplane. What should be the price of the ticket? If there are several airlines competing on this route, it should be close to $200. It cannot be less, at least in the long run; otherwise the airlines would lose money. It should be a bit more because airlines need to recoup their fixed costs. Let’s say this requires a 5 percent margin. Then the competitive price would be $210, and the average profit per passenger for this route would be $10.

  Now imagine that there is only one airline offering a direct flight. It would certainly try to charge more than $210. But how much more? It depends on how quickly it loses passengers when it raises its price. It can lose passengers in one of three ways. People could switch t
o another airline that offers connecting flights; they could decide to drive or to take the train; they could decide not to travel at all. The speed at which the airline loses customers when it raises its price is called the elasticity of demand.

  If the number of passengers decreases by 2 percent when the airline increases its prices by 1 percent, then we would say that the elasticity of demand is 2. This elasticity depends on the outside options of the passengers. If they can find cheap and convenient connecting flights, their elasticity of demand is high, and the monopoly airline will not be able to charge much more than $210 for the direct flight. If there are no convenient connecting flights, the elasticity of demand is low, and the monopoly airline can charge much more than $210 for the direct flight.

  Market Power and Welfare

  Now let’s extend the argument to show the relationship between market power and welfare, beginning with a competitive industry. In Figure 2.1a, prices are plotted on the vertical axis and quantities on the horizontal axis.

  The demand curve measures consumers’ willingness to pay. Imagine this is the market for a car, and you are ranking consumers from the most eager to the least eager to buy. The price pins down the marginal buyer—the person who is indifferent about buying the car at that price. As shown in Figure 2.1a, all of the consumers to the left of the marginal buyer would be willing to pay more, but of course they are happy to pay less. The ones to the right of the marginal consumer decide to pass. The price is just too high for them.

  You can also imagine that the good is chocolate, and there is only one consumer. She likes chocolate and would be willing to pay a relatively high price to get some chocolate instead of none. As the quantity goes up, her cravings are gratified, and her willingness to pay for an extra ounce goes down. From our perspective, it is equivalent to think of many consumers buying one unit of a good (such as one car) or one consumer buying several units of the good (ounces of chocolate, for example). Either example creates the same downward-sloping demand curve with which we can perform the same welfare analysis, so you can pick your favorite example. For simplicity, I will continue with the car metaphor.

  The next step is to figure out what the price is going to be. Clearly, that will depend on how firms compete to supply the goods. If the industry is competitive, the price must equal the marginal production cost—the price to build one extra car or to produce one extra ounce of chocolate. Why? Because if the price was above the cost, at least one firm would have an incentive to lower its price and attract new customers. Firms would keep undercutting each other until the price equals the cost. At that point they would break even and would not lower the price further. This is what we mean when we say that the industry is competitive. At that competitive price, we can figure out how many people want to buy by looking at the demand curve. We find the marginal buyer, and we can read the quantity on the horizontal axis. This is the competitive quantity.

  FIGURE 2.1  Industry equilibrium. (a) Competitive industry; (b) Industry with market power.

  The gray triangle in Figure 2.1a measures the consumers’ surplus. It should be relatively easy to see why. Each point on the demand curve shows the buyer’s willingness to pay; it measures how much she values the car. When her value is above the price, she earns a surplus, measured exactly by the distance between the demand curve and the price. The triangle is the sum of all the surpluses of all the buyers to the left of the marginal one. Naturally, the marginal buyer has a zero surplus, which is the same as saying that she is indifferent to buying and not buying the car at the current price.

  To keep things simple, in this example we have assumed that the marginal cost does not depend on the quantity produced: we have drawn a horizontal line for the marginal cost in Figure 2.1a. When the price equals the cost per car, firms make no extra profits, so there is nothing to count on their side. Consumer surplus is the correct measure of total welfare in this economy. You can imagine more realistic cases in which the marginal cost first decreases and then increases without changing the main points of our analysis.

  Figure 2.1b shows the same economy when firms have market power. The price is now above the marginal production cost. Fewer people buy cars, and the quantity is lower than in the competitive market. Consumer surplus is measured by the smaller triangle. You can see that it’s a lot smaller than the previous one. Our marginal buyer and a few to her left have been priced out, and everyone else is paying a higher price.

  The entire difference between the two triangles is not a welfare loss, however, because the light gray rectangle indicates the firms’ profits. These were zero in the competitive case, but now they are positive. These profits are not lost. They are paid to shareholders and recycled in the economy. Of course, the profits accrue to shareholders while the high prices are paid by consumers, and these are not necessarily the same people. As I was writing this book, I exchanged some emails with Howard Rosenthal, a political scientist at New York University. In one of them he wrote something that perfectly captures this issue: “When I look at my telecom bills here and what they were when I had a place in Paris, the difference is obvious. But I love my dividend checks from AT&T and Verizon.” We will have to think more about the distributive impact of higher prices. We will return to this theme later in the book, but for now the important point is that the efficiency (or welfare) loss is represented only by the small black triangle.

  Since competition is good for welfare, the obvious question is this: How do we move from Figure 2.1b to Figure 2.1a? What does it take to make an industry more competitive? And do we see gains for consumers? Let us study some specific industry cases.

  The Deregulation of Airlines and Telecoms

  In the Introduction we discussed the deregulation of airlines and telecoms. The Carter administration began the deregulation of airlines. Lower prices led to a sustained increase in the number of passenger-miles flown each year. The “Baby Bells” were carved out of AT&T in 1984. As with the airlines, increased competition in long-distance service resulted in significant benefits for consumers. Prices of long-distance phone calls decreased dramatically. AT&T’s market share fell from above 80 percent in 1984 to about 50 percent in 1996 as more competitors entered the market.

  The deregulation of telecoms happened much later in Europe. The case of France provides a striking example of the virtue of competition. Free Mobile is the wireless service provider of the Iliad group, a telecom company founded by French entrepreneur Xavier Niel. It obtained its 4G license in 2011 and became a significant competitor for the three large incumbents. The impact was immediate. Until 2011, French consumers had to pay between €45 and €65 per month for their smartphone plans, with limited data and a few hours of talk time. Free Mobile offered unlimited talk, unlimited SMS and MMS messages, and unlimited data for €20. The number of Free Mobile clients grew quickly, from about 2.6 million in the first quarter of 2012 to over 8.6 million in the first quarter of 2014. Free’s current market share is around 20 percent, and it aims to achieve 25 percent.

  The benefits to consumers spread far and wide: incumbents Orange, SFR, and Bouygues reacted by launching their own discount brands and by offering €20 contracts as well. In about six months after the entry of Free Mobile, the price paid by French consumers had dropped by about 40 percent. Wireless services in France had been more expensive than in the US, but now they are much cheaper. Quality also improved, as anyone who has made calls in both countries can attest.

  These cases are relatively clear cut and easy to interpret: an action taken by the government (deregulation and an antitrust suit) leads to increased competition, and all the indicators move in the same direction. We observe lower prices, lower profits, and lower concentration, at least initially.

  There are other examples, however, in which our three indicators (concentration, prices, and profits) do not all move in the same direction. For instance, when search and transport costs decrease, consumers can more easily buy from low-cost producers, and concentration c
an increase even though the market is competitive. Chad Syverson (2004), an economist at the University of Chicago, documents this effect in his study of plants that produce ready-mixed concrete. He finds that “when producers are densely clustered in a market, it is easier for consumers to switch between suppliers … Relatively inefficient producers find it more difficult to operate profitably as a result.” Competitive pressures force inefficient producers to exit, and the market share of efficient firms increases. In this case, therefore, competition leads to higher concentration.

  Efficient Concentration: Walmart in the 1990s

  Let us consider the case of retail trade, and specifically the expansion of Walmart. Walmart had a large impact on the US retail sector in the 1990s, and improvements in the retail sector had a large impact on overall economic growth during that period, driving as much as one-third of the improvement in economic efficiency between the first and second halves of that decade.a

  FIGURE 2.2  The growth of Walmart

  Figure 2.2 shows the market share and profit margin of Walmart. The market share is simply the ratio of Walmart’s sales (revenues) over the total sales of the retail sector. Note that we need to define what we mean by “retail sector.” The graph shows the market share of Walmart within the category of general merchandise stores. If you want to understand how economists classify firms into industries, you can look at the first section in the Appendix, where I explain everything you always wanted to know about industry codes but were afraid to ask. Walmart’s market share grew dramatically in the 1990s, from less than 5 percent to almost 60 percent. The profit margin is defined as profits over sales. If the profit margin is 5 percent, it means that when Walmart sells $1 of goods, it makes a profit of 5 cents. Walmart’s margin went down a little bit over the period, from 6–7 percent to 4–5 percent. The slight decline in margin can be explained by the astonishing increase in its revenues. The profit margin is an average, and the expansion presumably entailed increasing sales of relatively lower margin products.