The Great Reversal

by Thomas Philippon

  Alexander Fouirnaies and Andrew B. Hall (2014) get around the issue in a smart way. They estimate the incumbency advantage by comparing, at the district level, the campaign contributions to the party that barely lost versus the party that barely won the previous election. They can then tease out the causal impact of campaign contributions on the likelihood of winning the next election. They find that campaign contributions explain a large part of the incumbency advantage, and that interest groups motivated by access to the politician account for two-thirds of the financial advantage of the incumbent.

  How much do political connections matter for businesses? To document the link between money and politics, you need really smart and creative researchers. Seema Jayachandran (2006) documented a clear “Jeffords effect” on US firms. In May 2001, Senator James Jeffords left the Republican Party and tipped control of the US Senate to the Democrats. Jayachandran focuses on the consequences of this switch. Firms more closely aligned with the Republican Party lost, and firms aligned with the Democratic Party gained. She measures alignment using a firm’s soft-money donations to the national parties. She uses stock market prices to assess how valuable politicians are to a firm. She estimates that an additional $100,000 donated to the Republican Party in the previous election cycle is associated with a .33 percent lower stock return during the event window—that is, the week of Jeffords’s switch.

  FIGURE 10.4  Contributions by industry sector to the Republican Party. FIRE = finance, insurance, and real estate; TCU = transportation, communications, and utilities

  Political connections are therefore clearly valuable to firms. But politics is notoriously risky. In this context, we would expect firms to hedge their bets. And they do. Figure 10.4 shows that industries give to both parties, though most of them give a somewhat higher fraction to Republicans.

  A Less Prominent Role in Europe

  Money seeks to influence politics in every country. In France, despite fairly strong campaign finance laws, Yasmine Bekkouche and Julia Cagé (2018) find that campaign donations do influence election results. They collected data on 40,000 candidates from four municipal and five parliamentary elections. The level and evolution of spending is quite different from what we have discussed in the US. Following changes in campaign finance laws in the 1990s—lower spending limits and prohibition of corporate contributions—there was a decrease in parliamentary election spending from around €22,000 per candidate in 1993 to €10,000 in 2007.

  Within an election cycle, however, there is still a tight connection between spending and votes, which brings us to the endogeneity issue. Perhaps candidates who are likely to win are also better at fundraising. Perhaps people like to give to winners. We cannot conclude that money buys votes simply by looking at the correlation. That is where research ingenuity really kicks in. Bekkouche and Cagé note that, between 1993 and 1997, France enacted a law prohibiting contributions from legal entities, such as corporations and unions. The law was applied for the first time in the 1997 legislative elections. The law affected mostly those candidates who previously relied on private donations from legal entities. Bekkouche and Cagé estimate that an additional euro received from legal entities in 1993 is associated with a 0.46 euro decrease in total revenues between 1993 and 1997. In other words, these candidates could replace only about half of the lost revenues.

  Here is the really cool part. We agree that we cannot use the actual spending pattern to show that money can buy votes, because of reverse causality and omitted variable bias. But we can use the predicted change in revenues triggered by the change in legislation. That predicted change in revenue is not caused by the ability of the candidate or by the expectation of winning. The predicted loss in revenues can be used to tease out the causal role of money in winning elections. They find a significant causal effect of spending on vote shares in the 1997 election. The price of a vote is about €10 for the legislative elections.

  This discussion could be expanded to most other European countries. Money influences politics everywhere. What sets the US apart is therefore not that money seeks to influence politics; it is rather the scale of the money involved that is exceptional. If differences in lobbying expenditures between the US and the EU are large (a factor of two, or three for corporate lobbying, as we saw in Chapter 9), differences in campaign contributions are staggering. Figure 10.5 shows total campaign contributions for federal elections in the US and total campaign expenditures for several European countries, normalized by GDP. The sample of European countries is primarily based on EU Parliament in 2015 and was chosen to be representative of the European economy. Campaign contributions in the US are fifty times larger than those in most European countries.

  FIGURE 10.5  Total campaign expenditures divided by GDP. Data sources: US, Center for Responsive Politics; EU, EU Parliament (2015). For Germany, see Bundestags-Drucksache (2013).

  The evidence is once again consistent with the model developed in Chapter 8. Europe has so far avoided the outsized role of money in politics that we observe in the US. Money in politics spills over to regulatory agencies. Federal Trade Commission and Department of Justice officials are likely to be influenced by elected politicians, or at the very least, elected politicians will attempt to influence the process. For instance, upon initiating its investigation of Google, the FTC received a number of letters from members of the US Congress, including at least one encouraging that agency to desist, noting the ability of Congress to limit the FTC’s power. Members of the European Parliament would be unlikely to write such letters, and even if they did, their efforts would not be very influential. The European Commission’s Directorate-General for Competition (DG Comp) is entirely independent from actions taken by the European Parliament.

  State Politics in the United States

  Let us now turn to state politics in the US. This research is particularly useful because there are many more elections to consider, and they do not all happen at the same time, thereby giving us a chance to rule out confounding factors.

  State regulators play an important role in the US economy. State attorneys general are in charge of competition policy. The attorney general serves as the chief legal adviser and chief law enforcement officer for the state government and is empowered to prosecute violations of state law. The National Association of Attorneys General (NAAG) facilitates interaction among attorneys general and collects data that we will use below.

  Campaign finance is also regulated at the state level. Robert J. Huckshorn (1985) identifies five primary types of regulation across states: restriction of the source of campaign contributions; restriction of the size of contributions; restriction of the size of political expenditures; disclosure laws; and public campaign financing laws.

  There is evidence that unlimited contributions help mostly incumbents. Thomas Stratmann and Francisco J. Aparicio-Castillo (2007) use changes in campaign finance regulations across states to show that races are more competitive in states with stricter limits on campaign contributions. Specifically, they find that limits on contributions reduce the advantage of incumbents. Keith E. Hamm and Robert E. Hogan (2008) find that contribution limits tend to reduce incumbency advantage and make electoral races more competitive. Timothy Besley and Anne Case (2003) focus on limits to corporate contributions and find that states with tighter restrictions have higher voter turnout and a higher fraction of women and Democrats holding positions in the legislature. Robert Feinberg and Kara Reynolds (2010) study the determinants of state-level antitrust activity. They find that states with larger economies and larger government expenditures file more antitrust cases and that antitrust activity increases during periods of high unemployment. They also find that state attorneys general who are appointed to their position file fewer antitrust actions than attorneys general who are elected.

  FIGURE 10.6  The type and number of enforcement cases with state attorneys general as plaintiffs. Data source: National Association of Attorneys General (NAAG) State Antitrust Litigation
Database

  Germán Gutiérrez and I have tested the effects of political expenditures on enforcement at the state level. We gather a list of antitrust enforcement cases initiated by state attorneys general from the Antitrust Multistate Litigation Database. We obtain campaign contributions for state elections from the Campaign Finance Institute. Case data is available since 1990, but contributions are available only after 2000. Figure 10.6 shows that state-level enforcement has decreased since the 1990s, just like federal enforcement. The decrease is particularly pronounced for nonmerger cases involving monopolization or collusion.

  Figure 10.7 plots nonmerger antitrust enforcement cases at the state level against total campaign contributions, which have nearly doubled since 2003. We use the four-year moving average because contributions exhibit substantial seasonality—increasing in years with gubernatorial elections.

  We then ask if these two trends are related. To do so, we check to see if state campaign contributions predict the number of enforcement cases for each state election cycle. The nice feature of state-level panel data is that we can net out the effects of election cycles and persistent state heterogeneity (we include state and election-cycle fixed effects). We can also control for a state’s economic conditions (growth and unemployment).

  FIGURE 10.7  State political contributions and nonmerger antitrust cases. Four-year moving average contributions control for the seasonality of election cycles. Data sources: Case data, NAAG State Antitrust Litigation Database; state campaign contributions, Campaign Finance Institute

  We find that high contributions in a state’s election cycle predict significantly fewer nonmerger enforcement cases in the following years. Broadly speaking, then, we find that companies strategically use campaign finance contributions across states to shield themselves from future enforcement cases.

  Dark Money, Charitable Foundations, TV Commercials, and Revolving Doors

  There are significant gaps in our ability to track political expenditures. We do not know the source of the money for a growing fraction of political spending. Political nonprofits do not have to disclose their donors. They can if they wish, but if they do not, no one knows where the money comes from. In theory, super PACs are supposed to disclose their donors. The problem is that they can accept donations from political nonprofits that have no obligation for disclosure. Spending without disclosure of donors has increased dramatically since 2008, and since 2012, has been more than $100 million per election cycle.

  The nonpartisan, nonprofit advocacy organization Issue One has traced most of the increase in “dark money” to spending by only fifteen groups.e These groups include labor unions, corporations, megadonors, and other special interest groups. As the report explains, “dark money comes into politics in many forms—from opaque limited liability companies (LLCs) to secretive social welfare organizations and trade associations.” Issue One provides a few striking examples. In Utah in 2011 “an innocuously named group called Freedom Path began airing ads” promoting Sen. Orrin Hatch, the incumbent, and opposing his primary challenger, state senator Dan Liljenquist. It was only in November 2012—months after the primary election and weeks after Hatch had won the general election—that a public document revealed that “a trade association called Pharmaceutical Research and Manufacturers of America (PhRMA), the nation’s pharmaceutical drug lobby, had provided nearly 90 percent of Freedom Path’s initial funding in 2011.” And, as one might have guessed, Senator Hatch enjoyed a friendly relationship with PhRMA. Voters in Utah did not have access to this information because, following the Supreme Court’s Citizens United decision discussed earlier, Freedom Path was formed as a nonprofit “social welfare” organization under Section 501(c)(4) of the US tax code—allowing its donors to remain hidden. In a more recent case, “one month before the 2017 special election to fill the Senate seat vacated by incumbent Republican Sen. Jeff Sessions … a super PAC called Highway 31 popped up in Alabama and began spending more than $4 million to boost Democratic Senate candidate Doug Jones.” A legal loophole allowed Highway 31 to avoid disclosing that it was controlled by groups aligned with the Democratic Party until a month after the election.

  Dark money is both prevalent and difficult to trace. But it’s when the task gets harder that great researchers stand out. Marianne Bertrand, Matilde Bombardini, Raymond Fisman, and Francesco Trebbi have managed to shed light on some of the dark money flows in a 2018 study. They show how firms use their philanthropic foundations to influence politicians. Bertrand and her co-authors find that these foundations give larger grants to charitable organizations located in congressional districts with a representative seated on a committee that matters for the firm. The pattern looks strikingly similar to that of publicly disclosed political action committee (PAC) spending. Furthermore, they show that a member of Congress leaving office leads to a short-term decline in charitable giving to his district, as well as in PAC spending. Charities directly linked to politicians exhibit similar patterns of political dependence. Bertrand and her co-authors show that firms deploy their charitable foundations as a form of tax-exempt influence seeking. Based on a straightforward model of political influence, they estimate that at least 7.2 percent of total US corporate charitable giving is politically motivated. This is almost three times larger than annual PAC contributions, and this is a conservative estimate. Charitable giving is not subject to formal disclosure requirements. It is therefore a form of political influence that goes mostly undetected by voters and shareholders, and which is directly subsidized by taxpayers.

  Firms are always looking for ways to influence politicians, and not only in the United States. Italian firms appear to be quite creative. In a fascinating paper, Stefano DellaVigna, Ruben Durante, Brian Knight, and Eliana La Ferrara (2014) show that Italian firms shifted their advertisement spending to benefit Silvio Berlusconi while he was in power. Berlusconi was in office three times between 1993 and 2009, and he maintained control of Italy’s major private television network, Mediaset, throughout. They find a significant pro-Mediaset bias in the allocation of advertising during Berlusconi’s political tenure, especially for companies in more regulated sectors. They estimate that Mediaset profits increased by one billion euros during this period and that regulated firms anticipated sizable returns on their political investment.

  Finally, there is the revolving door issue. Thomas Boggs, the famed lobbyist whom we encountered in Chapter 9, “helped pioneer the ‘revolving door’ culture of hiring former members of Congress and others with enough prestige to get the right people on the phone, fast,” noted the Washington Post in Boggs’s 2014 obituary. Just like lobbying, revolving doors can be good or bad. When you hear people complain about “professional politicians” or the lack of understanding of private sector challenges among civil servants, you are effectively hearing people asking for more revolving doors. Revolving doors are useful when they allow a better sharing of expertise and information. They are bad when they distort the incentives of regulators or lead to regulatory capture. In Washington, DC, there’s no shortage of examples of regulators leaving their government jobs to join a company that lobbies the very agency they worked for, or powerful executives of major firms taking jobs as top regulators of their former companies, often after receiving a generous payout on the way out the door.

  One of the most egregious revolving doors over the years has been at the Federal Communications Commission, which regulates, among other things, telephone, internet, and television services. Michael Powell, FCC chairman from 2001 to 2005, would go on to become CEO of the National Cable and Telecommunications Association (NCTA) in 2011. Jonathan Adelstein, an FCC commissioner from 2002 to 2009, left to take over PICA, a trade group representing wireless telecommunications firms. In 2011, FCC commissioner Meredith Baker left her job after serving only two years of a four-year term and jumped over to a top lobbying position at Comcast. In 2013, Barack Obama appointed Tom Wheeler, a former president of the NCTA and a former CEO of the Cellular Comm
unications & Internet Association, to be FCC chair. When Wheeler resigned, Donald Trump appointed Ajit Pai, a former Verizon employee and communications industry lawyer, to replace him.

  The movement of executives between regulatory agencies and the businesses they regulate is well documented in multiple industries. David Lucca, Amit Seru, and Francesco Trebbi (2014) study workers’ flows between US banking regulators (both federal and state) and the private sector. Workers’ flows within the private sector are large. Workers are constantly moving from job to job, losing a job and finding a new one. Between one-third and one-half of all new hires are workers moving from one firm to another. There are also significant flows between financial regulators and private firms. On average, each year, about 5 percent of regulators leave their job for the private sector and about the same number travel the other way. These total gross flows are thus around 10 percent annually. Total gross flows within the private sector are typically around 20 percent to 25 percent annually (for instance, in the Current Population Survey). The regulator-to-private-industry flows are thus between a third and a half of the private-industry-to-private-industry flows, but they have increased over time. Regulatory agencies face a retention challenge, especially for their most talented employees.

  The trouble with revolving doors is that they can lead to regulatory capture. Capture can be direct (quid pro quo) or intellectual (ideological). Haris Tabakovic and Thomas Wollmann (2018) find evidence of direct capture in the regulation of patents. Using detailed data from the US Patent and Trademark Office, they find that patent examiners grant more patents to the firms that later hire them or are likely to hire them. These firms also enjoy better intellectual property protection. This might be consistent with some form of information sharing, but Tabakovic and Wollmann also find that the extra patents granted are of lower quality, based on the observation that they are less frequently cited in subsequent patent applications.