The Great Reversal

by Thomas Philippon

  CHAPTER 9

  Lobbying

  Careful investment in a Washington lobbyist can yield enormous returns in the form of taxes avoided or regulations curbed—an odd negative sort of calculation, but one that forms the basis of the economics of lobbying.

  JEFFREY BIRNBAUM, THE LOBBYISTS

  BIG BIRD, the giant yellow Muppet beloved by generations of children on Sesame Street, may seem like an odd figure to mention at the outset of a chapter on political lobbying. But in the mid-1990s, Big Bird played a key role in a very public lobbying campaign that forced the most powerful politician in the US into an embarrassing retreat.

  After Republicans, led by Newt Gingrich, the firebrand Speaker of the House of Representatives, took over Congress in 1995, they targeted the Public Broadcasting Service for budget cuts. PBS produces, among other programs, Sesame Street. Gingrich and his fellow Republicans believed that the federal government had no business funding a media outlet, and suspected that PBS was indoctrinating people with a liberal worldview.

  The plan was to slash federal funding for PBS in the upcoming budget bill. But that was before a grassroots lobbying campaign was launched by PBS supporters, who were clever enough to cast the debate as a battle between Gingrich and Big Bird. With some careful nudging from PBS—which was technically barred from lobbying Congress itself—PBS viewers across the country practically buried their representatives in Washington with letters of support for Big Bird and, by extension, PBS.

  “It was a wonderful lesson in democracy,” an executive with one local station told the New York Times. “Here was ‘evil Newt’ trying to kill Big Bird. There was a public outcry and we were inundated with calls from viewers. We told them that if they cared about the issue, they should let their elected representatives know.”

  In the end, Gingrich was forced to back down. (In a mock ceremony before the second session of the 104th Congress, he even publicly swore not to “kill” Big Bird.) In the budget bill for fiscal 1997, Congress appropriated $300 million for PBS as far out as 2000, a 20 percent increase over the previous year’s funding.

  Normally, of course, lobbying efforts in Washington are rarely as easy to spot as the campaign to save Big Bird, and their results can be even more obscure. This makes large-scale data hard to collect, and research tricky.

  We’ve spent a lot of time analyzing data in previous chapters of this book—but at least it was good data. Aggregate economic quantities are complex to measure, but at least the numbers are relatively transparent. We can measure investment, employment, and incomes. We can relatively easily measure sales at the firm level. The study of prices across countries is more difficult, but it’s still a walk in the park compared to the study of lobbying and political influence over time, across industries, and between countries.

  To analyze the impact of lobbying on political outcomes is to enter the realm of bad data and wild guesses. When it comes to lobbying and campaign contributions—expenses to influence the political and regulatory process—we can really only measure the tip of the proverbial iceberg. An especially pernicious effect of lobbying and campaign contributions is that the lack of pristine data creates a tendency to let the data available dictate the questions we ask. In the normal course of research, you first choose a question that you deem interesting and important, and then you look for ways to answer it. This, however, can sometimes be terribly slow and frustrating, especially when there is no available data. It is also risky because you might never get the data you need to obtain a convincing answer. There is a natural tendency, then, to reverse the logic. You might decide to work with the data you already have, knowing full well it will not let you answer the more interesting question. At least it will let you answer something. This is the research equivalent of looking for your keys under the streetlight.a

  Fortunately, political economy is a field where we will encounter some of the most creative and pugnacious researchers in economics. Thanks to their work, we will be able to draw some striking conclusions about the role of lobbying and money in American politics.

  There is nothing inherently bad in lobbying. In fact, hiring a lobbyist is a right protected by the US Constitution. The First Amendment of the US Constitution protects free speech and guarantees the right “to petition the Government for a redress of grievances.” Political scientists John de Figueiredo and Brian Richter (2014) explain that

  One of the central tenets of representative democracy is the right of individuals, by themselves or in groups, to petition elected officials and the government. These petitions are designed to influence the opinions, policies, and votes of legislators and other government officials. One outgrowth of this right has been the creation and evolution of organized interest groups comprised of individuals, companies, and other organizations. These organized interests employ a variety of methods to influence government policies including campaign contributions, endorsements, grassroots campaigns, media campaigns, and lobbying.

  In a democracy, citizens have the right to petition their government, and we all agree that it is a good thing. But it also opens the door to lobbying by large corporate interests. As with most things in life, then, it is a matter of balance. Let us find out how this balance has changed over the past twenty years.

  Why Measuring the Impact of Lobbying Is an Uphill Battle

  Before discussing lobbying and campaign contributions, I should explain at the outset why it is difficult to measure the outcome of these contributions. Small legislative or regulatory changes—often buried in larger documents running to hundreds of pages—can be worth millions of dollars to a particular industry or interest group. But they can be virtually impossible for a nonexpert to identify and to understand. Harder still is detecting what things an effective lobbying effort persuades legislators to leave out of a bill entirely. And even when such changes can be uncovered, it is often impossible to definitively link them to specific lobbying efforts or campaign contributions.

  This explains the lack of consensus in the literature on the real effects of lobbying. It has led some commentators to doubt that lobbying matters at all. This strikes me as implausible, for at least four reasons.

  First, at a theoretical level, you would need to hold a rather strange view of the world to think that lobbying does not matter. As Marianne Bertrand and her co-authors (2018) explain: “At the intersection between the political and the economic spheres lies the lobbying industry. Trillions of dollars of public policy intervention, government procurement, and budgetary items are constantly thoroughly scrutinized, advocated, or opposed by representatives of special interests.” In order to deny the influence of lobbying, you would have to explain why businesses voluntarily spend substantial human and financial resources on something useless. The fact that lobbying has always taken place (and, if anything, seems to be increasing) is enough, under conditions of minimal rationality, to argue that it must matter. Put another way, if we hold the view that firms spend money on lobbying but it is useless, then we must recognize that this view is inconsistent with most of what we know about economics and human nature. It’s not strictly impossible, but I am not going to take this idea very seriously.

  Second, rent seeking is a zero-sum game, and zero-sum games are difficult to identify in the data. For instance, suppose firm A spends $100 to lobby for a regulatory change that would give it an advantage over firm B. Firm B then spends $200 to fight this change. Firm B prevails. What do we see? We see that they collectively spent $300 and that nothing has changed: the relative market shares, growth rates, and productivity of the two firms are the same as before. Any naive model would conclude that lobbying does not matter, even though it obviously does. This is not just a theoretical possibility. Frank R. Baumgartner, Jeffrey M. Berry, Marie Hojnacki, David C. Kimball, and Beth L. Leech (2009) conducted a long-term study of the lobbying efforts of 2,200 lobbyists. They found that both sides of the issues that they analyzed were able to muster roughly similar resources.

  Third
, rent seeking, even when it is legal, is not something that the parties involved want to advertise: neither the firms at the contributing end, nor the regulators and politicians at the receiving end. We should therefore expect the outcomes to be hidden. This creates a particularly difficult problem for researchers. Difficult does not mean impossible, however. Carnegie Mellon economist Karam Kang (2016) used a sophisticated model to estimate the returns from lobbying in the energy industry. She built a data set containing all federal energy legislation and lobbying activities by the energy sector during the 110th Congress. She then computed a game-theoretic model of lobbying. This means that she described a game between lobbyists and policy makers using mathematical equations. This allowed her to estimate the returns from lobbying expenditures. The average return is above 130 percent.

  Fourth, and finally, contributions are choices, not random inputs. In the language of econometrics, contributions are endogenous variables and the research is plagued by endogeneity and omitted-variable bias. Human beings have an unfortunate tendency to think ahead, at least some of the time. This makes my life as a researcher rather difficult.

  People Do Stuff for a Reason

  Endogeneity bias is a pervasive issue in economics and in social sciences more generally, so it’s worth taking a moment to explain what it is. In fact, once you think about it, you’ll see that is one of the two main differences between the natural sciences and social sciences, the other difference being the ease or difficulty of running controlled experiments. Let me illustrate with a few examples of endogeneity bias.

  Suppose you want to answer the following question: is seeing a doctor good for your health? We know the answer is yes, at least on average. (I am not talking about bloodletting in medieval medicine.) But how would you test this simple idea? Suppose I gave you data on doctors’ visits paired with data on health outcomes six months later. What would you find? Well, you would find that people who visit doctors are more likely to be sick or even dead six months later. Why? Because sick people are the ones who go to the doctor in the first place. Healthy individuals do not. Technically, we say that the decision to visit (or call) the doctor is an endogenous decision. This means that people make this decision for a reason. Therefore, you cannot treat the doctor’s visit as a random event as you would in an experiment in the natural sciences. How can you answer the question? You need to find changes in access to doctors that are independent of health status, such as the random opening or closing of a medical facility in the neighborhood. Even then, you need to be confident that the opening or closing is not itself driven by the average health of the population in the neighborhood.

  Another example involves the law of supply and demand. Suppose you want to know by how much people decrease their consumption of milk when the price of milk goes up. The two major factors in question are the price—how much does the grocery store charge for milk—and the level of demand—how much do milk consumers want? The problem is that in the real world, both of these factors are being determined simultaneously. The store owner adjusts prices according to demand, and consumer demand fluctuates with prices. Suppose the store owner can predict when the demand for milk will go up. On days when she expects the demand to be high, she might want to increase the price a little bit to increase her profit margin. What would you see in the data? You would see that people consume more milk when the price goes up, and you would be puzzled because that certainly does not look like a demand curve. The issue in that example is that price setting is endogenous and depends on expected demand.b

  Endogeneity is less of a problem in the physical sciences. Subatomic particles do not behave like human beings. Particles are not forward looking. They do not change their behavior because they expect something to happen. Humans do, fortunately for them, unfortunately for researchers. We will come back to these points, but for now it’s enough to say that these are tricky issues and that real ingenuity is needed to deal with them.

  How does this translate to lobbying and campaign finance? Let’s look at lobbying. Which firms or industries have an incentive to lobby? Well, precisely the ones that feel potentially threatened by new rules and regulations, or the ones that have something to hide, or the ones that have rents to protect. In later chapters we will study the giants of the internet economy, the GAFAMs (Google, Apple, Facebook, Amazon, and Microsoft). They have recently increased their lobbying expenditures. Do you think this is by accident, or do you think this is because they feel the risk of a regulatory backlash? Yes, this is a rhetorical question.

  In other words, firms that have an incentive to lobby are the ones that are most likely to be targeted. Just like with doctor’s visits or the price of milk, the endogeneity of the lobbying decision renders the simple correlation uninformative. The correlation between lobbying today and enforcement actions tomorrow might be zero or positive, and a naive model would conclude that lobbying does not work, or even that it backfires, just like a naive model would conclude that doctors’ visits are bad for health or that people drink more milk when its price goes up. We will discuss specific examples in the chapters on finance, health care, and internet firms. We will see how banks, asset managers, and internet giants all beef up their lobbying efforts precisely when they feel that some new regulations or new competitors might come their way.

  To deal with endogeneity, economists have learned to turn the twists of history into quasi-natural experiments. For instance, to understand the causal influence of political preferences on antitrust enforcement, Richard Baker, Carola Frydman, and Eric Hilt (2018) use the sudden accession of Theodore Roosevelt to the presidency after the assassination of President William McKinley. McKinley presided over the largest merger wave in American history. He had no interest in antitrust and no intention of putting limits on mergers. Teddy Roosevelt had very different opinions. Firms with greater vulnerability to antitrust enforcement lost more after the assassination. The researchers conclude that “the transition from McKinley to Roosevelt caused one of the most significant changes in antitrust enforcement of the Gilded Age—not from new legislation, but from a change in the approach taken to the enforcement of existing law.”

  Why Lobbying Creates Inefficiencies

  As we have noted, lobbying is not necessarily bad. There are basically two views of lobbying, one benign or even beneficial, and one negative. The benign view of lobbying is that it allows the sharing of relevant information between businesses, regulators, and politicians. Technology and tastes are always changing, and it is difficult for policy makers to keep up. It is difficult to know when “laissez-faire” is appropriate or when regulations are preferable. It is difficult to figure out what is important and what does not matter. In this context, lobbies can be beneficial. They have a strong incentive to provide relevant—even if biased—information. At the very least, they show that some people and businesses care about a particular issue.

  The alternative perspective holds that lobbying is essentially rent-seeking. That is the view articulated by economists Gene M. Grossman and Elhanan Helpman (1994, 2001). Simply put, businesses lobby to protect their rents. This explains, for instance, the steel and aluminum tariffs imposed by the Trump administration in 2018. Lobbying spending by large American steel companies increased by about 20 percent between 2017 and 2018. They successfully pushed the Trump administration to impose tariffs, which led to higher prices and higher profits. In the case of aluminum, a midsize company with fewer than 2,000 workers, Century Aluminum, was at the center of the lobbying effort. As Business Week reported on September 27, 2018, “what wasn’t mentioned was that Century’s biggest shareholder is Glencore Plc, the Swiss trading company that is the biggest buyer and seller of commodities in the world … While Century was lobbying the Trump administration, Glencore, along with a handful of other commodity trading companies, was stockpiling record amounts of foreign aluminum in the U.S.—the idea being, if tariffs were announced, prices would rise, and all that cheap foreign metal would suddenly become mo
re valuable. And that’s what happened.”

  Rent-seeking causes loss of wealth in two ways. First, the direct expenditures that interest groups spend on lobbying could be used for productive work, rather than zero sum games. The second loss is the policies themselves. The policies advocated by lobbyists are rarely efficient. They do not take the form of simple transfers or lump-sum taxes. Consider, for instance, the regulation of entry. Imagine a world where entrants pay a lump-sum tax to compensate for the disruption and harm they inflict on incumbents. In that world, there would be no indirect loss through inefficiency. New firms would still enter and bring benefits in terms of price competition and innovation. The entrant would have to write a check to the incumbent, but that would be a transfer, not a deadweight loss. Some criminal enterprises follow these rules because they understand that it’s efficient to let other businesses operate and “tax” them. It is also the policy advocated by many economists when entrants destroy the value of existing assets, such as Uber with respect to taxi medallions.

  Yet we rarely observe this kind of outcome in our economies because incumbent rents are usually not legitimate. The transfer would be too transparent and would cause an outcry. It would also not be enforceable ex post. The entrant could simply refuse to pay after it enters. As a result, incumbents typically resort to inefficient ways to protect their rents. Oftentimes it means blocking entry altogether. That creates a large inefficiency.

  What You Know or Whom You Know?

  It is difficult to assess how much of the lobbying that we observe reflects the useful sharing of information, and how much is rent seeking. We can, however, answer a related question that sheds light on the issue. We can ask whether lobbyists provide specific information to members of Congress or whether they provide special interests with privileged access to politicians. This later view is consistent with what insiders believe, as Conor McGrath (2006) reports: “there are three important things to know about lobbying: contacts, contacts, contacts.”