Some evidence clearly suggests that whom you know matters. Jordi Blanes i Vidal, Mirko Draca, and Christian Fons-Rosen (2012), for example, show that lobbyists who used to work as senatorial aides experience a significant drop in revenue when their senators leave office. Also consistent with this interpretation is the fact that the big lobbying firms, such as Squire Patton Boggs or Cassidy and Associates, are not exactly famous for their technical expertise.
Thomas Hale Boggs Jr., whose name is associated with the Patton Boggs firm (now Squire Patton Boggs), is a prime example of connections being the key to success in the lobbying business. His parents both served in Congress. As the Washington Post wrote in his obituary (September 15, 2014), “Young ‘Tommy’ Boggs’s first job in Washington was operating the private elevator for then-House Speaker Sam Rayburn, the Texas Democrat.” Able to get virtually anyone in Washington on the phone on demand, Boggs held court at the famed Palm restaurant in DC and used his vast web of connections to lobby for clients of all descriptions. He helped the American Bankers Association in its effort to overturn the onerous regulations of the Glass-Steagall Act, was instrumental in rewriting a massive telecommunications bill in 1996, and helped a string of foreign politicians, some of them rather unsavory, secure the assistance of the US government. To be clear, Boggs was no expert in banking, telecommunications, or foreign relations. But he always knew the right people to call, the effective place to apply political pressure, and the value of a well-placed contribution.
It’s pretty clear that whom you know is important, but that does not rule out that what you know might also be important.
Until recently we did not know much about the relative importance of these two effects. Thanks to the brilliant work of Marianne Bertrand, Matilde Bombardini, and Francesco Trebbi, however, we now have a sharper picture. They show that both channels (what you know and whom you know) matter, but that whom you know matters more. They show that lobbyists follow politicians to whom they were initially connected when those politicians receive new committee assignments. For instance, “a lobbyist who is connected to a legislator whose committee assignment includes health care in one Congress is more likely to cover defense-related issues in the next Congress if the legislator he or she is connected to is reassigned to defense in the next Congress.”
They also find some support for the expertise view. They show that there is a group of experts whom even politicians of opposite political affiliations listen to. Politicians often maintain relationships with lobbyists of the same political orientation, but they are more likely to cross the aisle when talking to experts, as one would expect if they are trying to obtain accurate information.
Bertrand and her colleagues are thus able to document both effects. On balance, however, they find a significantly larger monetary premium for connections than for expertise. And the expertise that seems to matter the most is knowledge of a politician’s constituency. Knowledge of specific policy issues appears less important.
Lobbying for Fiscal Privilege
Finally, a lot of lobbying directly targets government purchases, transfers, and taxes. Beth L. Leech, Frank R. Baumgartner, Timothy M. La Pira, and Nicholas A. Semanko (2005) show that more lobbying resources are spent on issues and agencies with larger budgets. Trade associations often lobby for lower taxes. Lobbying for lower taxes is fundamentally inefficient because tax breaks create distortions in the allocations of economic resources, and because someone else must then pay these taxes. The economists Tanida Arayavechkit, Felipe E. Saffie, and Minchul Shin (2014) show that when firms lobby for capital-based tax benefits they become too large. They use lobbying and firm-level data from the US to document that firms that lobby are larger, more capital intensive, and enjoy lower effective tax rates. Their marginal product of capital is lower than that of firms that do not lobby. They argue that lobbying firms overaccumulate capital by 5.5 percent on average.
You might think that lower taxes can have beneficial incentive effects. That is true, but that is almost never the outcome of lobbying. When economists advocate for lower taxes, we mean lower marginal tax rates on as broad a base as possible. The tax breaks obtained by lobbyists take the form of loopholes and rarely improve investment and hiring decisions.
Lobbying Both Sides of the Atlantic
We have explained why lobbying is likely to be inefficient, and why it is difficult to measure. Let us now come back to our comparison of the US and Europe. The theory presented in Chapter 8 predicts higher lobbying expenditures in the US than in the EU. Is that true?
Figure 9.1 shows total lobbying expenditures directed at the US federal government and at EU institutions. Lobbying expenditures in the US appear to be more than twice as large as in Europe. Moreover, the share of lobbying done by business, lawyers, and lobbyists is higher in the US (87 percent) than in Europe (70 percent). Several caveats are warranted with these data but the differences are so large that even if we are 20 percent off, it would not change our main conclusions. The most important caveat is that many lobbying expenditures go unreported, and it is difficult to know which way the bias goes. In the US, nonprofit entities play an active role in the lobbying process. Financial ties between firms and nonprofit entities are legal and tax exempt, but difficult to trace. Marianne Bertrand and co-authors (2018) find that corporations strategically deploy charitable grants to induce nonprofit grantees to make comments that favor their benefactors. In Europe, there is a long tradition of dark lobbying. It is common wisdom among Italian political scientists and economists that there is an entire lobbying industry in Italy that goes completely unreported, with wealthy individuals operating networks of connections while never being mentioned in the press. This bias might be smaller at the EU level, but it’s impossible to know for sure.
Figure 9.1 is consistent with the literature on lobbying. In their survey, John M. de Figueiredo and Brian K. Richter (2014) document four main empirical regularities about lobbying. The first regularity is that lobbying is pervasive in the US and in other developed countries. Lobbying expenditures at the federal level in the US are several times larger than campaign contributions by political action committees (PACs). “In 2012, organized interest groups spent $3.5 billion annually to lobby the federal government, whereas interest groups’ PACs, super PACs, and 527 organizations spent approximately $1.55 billion over the two-year 2011 to 2012 election cycle (or approximately $750 million annually) on campaign contributions.” If you do not know exactly what PACs and super PACs are, hold on until Chapter 10, which will tell you everything you need to know about campaign finance.
FIGURE 9.1 Lobbying expenditures in US and EU. See caveats for EU lobbying totals in the text. US business sector includes agribusiness, electronics, construction, defense, energy, finance, insurance, real estate, health, lawyers and lobbyists, misc. business, and transportation. EU business sector includes professional consultancies / law firms / self-employed consultants, and in-house lobbyists and trade / business / professional associations. Data sources: US, Center for Responsive Politics and Federal Lobbying Disclosure Act Database; EU, LobbyFacts.eu and the EU Transparency Register
The second regularity is that “Corporations and trade associations comprise the vast majority of lobbying expenditures by interest groups.” Indeed, we can see that fact in Figure 9.1. In contrast, issue-ideology membership groups represent 2 percent and 7 percent of lobbying expenditures at the federal and state levels, respectively.
The third regularity is that “large corporations … are more likely to lobby independently than are smaller groups,” while “small interest groups are more likely to lobby using only trade associations.” This, of course, is exactly what the theory of collective action would predict. Mathilde Bombardini and Francesco Trebbi (2011) consider the role of trade associations. These associations appear to be more effective than individual firms at influencing policies.
The fourth empirical regularity is that lobbying increases when the stakes are higher and th
e issues more salient. This is precisely what makes the empirical research challenging: it means that lobbying is endogenous and therefore that we cannot hope to estimate the impact of lobbying with naive statistical models and correlations.
How Skewed Are You?
Figure 9.2 looks at the lobbying and campaign finance activities of firms that belong to the S&P 1500 (roughly, the 1,500 largest firms in the US). We see that the fraction of firms that make positive campaign finance contributions or have positive lobbying expenses has grown over time. The fraction of S&P 1500 firms that engage in lobbying has increased from about 33 percent to about 42 percent.
Political activism by US firms is thus becoming more pervasive. At the same time, the distribution of campaign finance contributions and lobbying expenses has been and remains very skewed. Skewness is an interesting word that means different things to different people. When we say that an argument is skewed, we mean that it is biased, unfair, or misleading.
FIGURE 9.2 Fraction of politically active firms in S&P 1500
In statistics, we talk about a skewed distribution to describe an asymmetry around the mean. Symmetric data have a skewness of approximately zero. The normal distribution (the nicely bell-shaped curve) has a skewness of exactly zero. A distribution is skewed to the right when there is a long right tail of large outcomes.
Table 9.1 describes the distributions of firms’ sales, campaign finance contributions, and lobbying expenses. The distribution of the logarithm of sales across firms is (positively) skewed. Its skewness coefficient is 0.23. This is a well-known fact; the firm-size distribution has a fat right tail. This means that large firms play an outsized role in the economy. Table 9.1 also shows the elasticity of campaign finance contributions and lobbying expenses relative to sales. For campaign contributions, the elasticity is 0.63. This means that, on average, when the revenues of a firm increase by 10 percent, its campaign contributions increase by 6.3 percent. For lobbying, the elasticity is 0.67.
Given that the distribution of firms’ revenues is skewed and that contributions increase with revenues, we expect a high concentration of lobbying expenses and campaign contributions. Table 9.1 shows that this is indeed the case by looking at concentration ratios. Even if we only look within the S&P 1500, so only within large firms, we see that the top fifty firms account for 42 percent of sales but 49 percent of campaign finance and 54 percent of lobbying. If we group them by industries, and we look at the CR4 by industry, it is 52 percent versus 65 percent and 68 percent. Among the very large firms, the super-large therefore play an outsized role. But these comparisons among very large firms understate the actual skewness. If we consider all the firms in the economy, small and large, the average industry CR4 for sales is 15 percent. On average, the top four firms control 15 percent of the revenues in their industries. But they account for 35 percent of campaign finance contributions and 45 percent of lobbying expenditures. In other words, lobbying expenditures are three times more concentrated than revenues, which are themselves already fairly concentrated. This means that large firms play an even more outsized role in the political system than they do in the economy itself.
TABLE 9.1
Skewness of Lobbying and Campaign Finance Contributions by Firm Size
(logarithm of)
Among S&P 1500 firms
All firms
Skewness &
elasticities
CR50
Industry CR4
Industry CR4
Sales
0.23 (skew.)
42%
52%
15%
Campaign finance
0.63 (elas.)
49%
65%
35%
Lobbying
0.67 (elas.)
54%
68%
45%
The elasticities of campaign and lobbying expenses to sales are computed by regressing log(expenses) on log(sales) for expenses above $10,000 and controlling for year fixed effects. Source: Compustat and OpenSecrets.com
Gutiérrez and I show that this fact is true in essentially all industries. The CR4 for lobbying is almost always much larger than the CR4 for sales. Our data show that, in nearly all industries, the role of large firms in politics and lobbying is even larger than would be suggested by their economic size alone.
Let us go back to the comparison with Europe. How confident are we that lobbying is indeed a larger factor in the US? EU lobbying might be underestimated; joining the European Commission’s Transparency Register is not mandatory. However, we have seen that lobbying expenditures are extremely skewed, so what really matters is to capture lobbying by large firms. For these firms, as far as we can tell, the data seem reliable.c
FIGURE 9.3 Distribution of large lobbying firms in the EU and in the US. Only firms are included—no trade associations or nonbusinesses. EU bunching is a result of how these data were processed (reporting in bins). Data sources: US, Center for Responsive Politics; EU, LobbyFacts.eu
One may also worry that our measures of lobbying toward EU institutions is lower because firms must lobby their individual countries as well. But this is also true in the US. In fact, according to FollowTheMoney.org, a website of the Campaign Finance Institute, total lobbying expenditures for only twenty states in the US (which account for 58 percent of US GDP) totaled $1.43 billion in 2016—nearly as much as total lobbying to the EU.
We can perform a more precise comparison using firm-level data. Figure 9.3 considers the top 1,000 lobbying firms in the EU and in the US. The shapes of the curves are similar for both entities, suggesting that qualitatively similar economic forces are at play. The US curve, however, is higher, which indicates that the propensity to spend on lobbying is materially higher in the US. The same results hold if we control for sector fixed effects. Konstantinos Dellis and David Sondermann (2017) estimate an elasticity of lobbying expenditures to log-sales of 0.15 in the EU. Using a sample of US firms from Compustat, we obtain an elasticity more than four times larger (0.620) in the US. The punch line is clear. Large firms in the US spend a lot more on lobbying than large firms in the EU, and this explains the large differences that we observe in the aggregate.
Do Lobbyists Succeed?
Our final question is also the most difficult to answer for the reasons explained earlier: the decision to lobby is highly endogenous. The big tech companies beefed up their lobbying efforts precisely when they started to hear complaints about their size and behavior, which meant they were more likely to be investigated precisely when they began to lobby more.
Most of the existing research on lobbying has focused on legislative outcomes: either votes or the composition of committees. The study of regulations and the influence on regulators is more recent. Regulators are often appointed and reappointed by legislators, and their budgets are usually determined by the votes of legislators. Firms want to influence regulators to increase their profits, and regulators might want to please legislators to increase their chance of reappointment and the size of their budgets.
Rui J. P. de Figueiredo Jr. and Geoff Edwards (2007) analyze lobbying by telecom firms toward state public utility commissions. They find that firms with larger contributions receive favorable regulatory decisions. Guy Holburn and Richard Vanden Bergh (2014) study mergers and acquisitions in the electrical utility industry. They find that firms contribute significantly more in the twelve-month period before the relevant regulatory decision than in other periods. This shows that firms attempt to strategically contribute, but we do not know if they actually succeed in influencing the decisions of regulators.
FIGURE 9.4 Contribution of industries to aggregate lobbying expenditures, 1999–2014
Figure 9.4 breaks out the aggregate expenditures on lobbying among industries. The finance industry is among the largest overall contributors, followed closely by durable and nondurable manufacturing and other services. Lobbying intensity also varies a lot across industries. Some industries (for exa
mple, finance) spend a much larger fraction of their gross income on lobbying than other industries (such as trade).
At the same time, some industries are targeted with a high number of antitrust cases, whereas some only receive a few. We focus here on nonmerger cases because the literature has already studied mergers in detail. Figure 9.5 shows the distribution of nonmerger cases among industries. The number of cases varies a lot and, as one would expect, industries with more cases (for example, durable and nondurable manufacturing) lobby more. This highlights the reverse causality issue.
FIGURE 9.5 Number of cases brought against industries, 1996–2016
The first thing we can do to get around the reverse causality issue is to look at changes in lobbying over time. We can ask whether increases in lobbying lead to decreases in the number of cases. We find that the answer is yes, but the significance is weak. This naive estimate suggests that a doubling of lobbying expenditures to the DoJ and FTC reduces the number of cases in a given industry by about 4 percent. On the other hand, we know that this estimate is biased downward by endogeneity. We just don’t know by how much.
We can do more if we use Europe to shed light on the US. Gutiérrez and I look at nonmerger antitrust cases and lobbying across industries (Gutiérrez and Philippon, 2018a). We find that there are more nonmerger cases in industries that are profitable and concentrated, as expected. Working with European data, however, gave us another idea. We can use EU cases as a measure of the unobserved “danger” that the industry faces, such as the threat of regulatory scrutiny. Let us assume for now that (1) lobbying in the US does not affect EU regulators, and (2) EU and US regulators have the same information about industry dynamics (technology, consumer taste, potential for bad behavior by large players, and so on).
The Great Reversal Page 18