Executives in Politics

1. Introduction

On November 8, 2016, Donald J. Trump won the U.S. Presidency. Although his election was unusual in many respects, Trump is just one of several examples of corporate executives running for political office. Other examples include Michael Bloomberg, a cofounder of Bloomberg L.P., who unsuccessfully ran for U.S. President in 2020, and Jon Corzine, a former chief executive officer (CEO) of Goldman Sachs, who in 2000 was elected U.S. Senator and in 2005 became the governor of New Jersey.¹ The growing prominence of business leaders in government has motivated public debate about their effect on policy and raised legitimate concerns that business politicians may unduly shift the balance of power toward corporate interests. To examine whether these concerns are justified, we assemble a novel data set on corporate executives’ involvement in U.S. politics and analyze their impact on firms, industries, and legislative outcomes over the last 40 years.

We start by documenting recent trends in the number of corporate executives elected to federal office. On average, there are 72 business politicians holding federal elected office in any given election cycle between 1980 and 2000, and this average increases to 100 per election cycle in the period from 2002 to 2018. In fact, after staying relatively constant at approximately 13% between 1980 and 2000, the share of business politicians among federal office holders goes up in every election after 2000 and rises above 22% by 2018 (Figure 1).²

One factor that may have contributed to the increase in the number of business politicians is the rising cost of political campaigns (Figure 1). When campaigns are expensive, candidates who can either self-fund themselves or have access to a network of wealthy donors are likely to have an advantage. Although we cannot establish a causal link between campaign finance and the rise of executives in politics, we present several pieces of evidence suggesting that business politicians have a fundraising advantage over their opponents. First, we find that business politicians are more likely to self-fund their campaigns. In fact, 24% of business politicians contribute at least $1 million of their personal funds to their campaigns compared with 4% of nonbusiness politicians. Second, business politicians outraise their opponents early on in the campaign, which prior research suggests is important for electoral success (Bonica 2017). Finally, we show that some of business politicians’ fundraising advantage comes from their firms, which donate six times more money to their executives running for office than to other political candidates.

In the next set of tests, we examine how the election of business politicians affects their firms, industries, and U.S. legislation more generally. We start by documenting that the election of corporate executives generates material firm-value benefits. Firms whose executives win federal elections experience positive abnormal stock returns of 1.4% within three days of such elections, which translates into a market value increase of approximately $200 million for an average firm in the sample. Furthermore, we find evidence of positive industry spillovers: when a business politician is elected, his/her firm’s industry peers also experience positive abnormal stock returns.

Next, we analyze business politicians’ legislative behavior and their impact on aggregate U.S. policy. Identifying this impact is challenging because the voting behavior of business politicians may reflect the underlying preferences of their constituencies. To separate the policy impact of business politicians from the underlying preferences of their electorate, we perform two sets of tests. First, we compare the voting records of business politicians to the voting records of their predecessors and successors, which enables us to contrast the voting behavior of business politicians with the aggregate long-run preferences of their constituencies. Second, we examine the voting records of politicians who win close elections to identify how business politicians change aggregate policies, even if voters do not necessarily demand those changes. To the extent that the outcomes of close elections are determined by chance, this latter set of tests allows us to separate the legislative impact of business politicians from the contaminating influence of aggregate voter preferences.

In both sets of tests, we find that business politicians are more likely, relative to nonbusiness politicians, to vote for legislation supported by corporate interests and against legislation supported by labor unions. The effects we document are economically important, even after controlling for party affiliation.³ The magnitude of the effect associated with being a business politician is about one-third of the unconditional averages of the different voting record measures that we use. Because this policy shift cannot be attributed entirely to the changes in the underlying voter preferences, our results provide a new perspective on how policy might drift in favor of business interests without voters demanding such a shift.

Our findings indicate that business politicians are systematically different from other politicians and that, once elected, business politicians shift the balance of power toward corporate interests. However, our results are notably silent on executives’ underlying motives for entering politics. One possibility is that executives run for office to advance the greater good but also have preferences and beliefs that are shaped by their prior business experience. For example, business politicians may be particularly receptive to the needs of the industries in which they previously worked. Our results on industry returns support this idea and suggest that business politicians may take actions favorable to their industry even if they do not specifically seek to benefit their firms or themselves personally. It is therefore possible that prior experience of elected politicians at least partly shapes their legislative behavior, suggesting that some political outcomes can be achieved only if voters elect a different set of politicians.

Our paper contributes to the large literature that studies interactions between firms and politicians. This literature explores different ways in which firms participate in politics and argues that such participation may be valuable to firms. For example, state-owned firms may receive more subsidies from the government in exchange for providing excess employment and paying above market wages (Shleifer and Vishny 2004, Bertrand et al. 2018). Non–state-owned firms can invest in political capital by establishing direct connections with legislators (Faccio 2006, Duchin and Sosyura 2012), lobbying (Bertrand et al. 2014, Kang 2016, Bombardini and Trebbi 2020), financing candidates’ political campaigns (Stratmann 1992, Cooper et al. 2010), and influencing political contributions of their employees (Babenko et al. 2020). These activities may create value for shareholders through greater government subsidies, favorable antitrust reviews, preferential allocation of federal contracts, less strict regulation, and lighter taxation.⁴

Another strand of literature examines firm-value implications of corporate officers’ entry into politics, with a particular focus on countries with weak institutions. Using data on publicly traded firms in 47 countries, Faccio (2006) shows that firms whose officers or large shareholders obtain public office experience positive abnormal stock returns, and this effect is stronger in highly corrupt countries. Gehlbach et al. (2010) show that corporate executives may enter politics to avoid the cost of lobbying elected officials in an environment with weak democratic institutions. Relatedly, Bunkanwanicha and Wiwattanakantang (2009) argue that weak institutions can allow business politicians to engage in rent-seeking activities. They study the entrance of business tycoons into politics around the 2001 election in Thailand and find that market valuations of tycoons’ firms increase substantially when the tycoons get elected. In contrast to this earlier literature, our focus is on the role of business politicians in an environment characterized by relatively strong democratic institutions. In fact, our paper is the first to document an increased involvement of corporate executives in U.S. politics and systematically study the consequences of such involvement not only for their firms but also for aggregate political outcomes.

The rest of this paper is organized as follows. Section 2 describes the data and documents the extent of corporate executives’ participation in politics over the last 40 years. Section 3 analyzes business politicians’ campaign fundraising. Section 4 investigates the effect of corporate executives’ electoral wins on their firms and industries. Section 5 examines the impact of business politicians on legislative outcomes. Section 6 concludes.

2. Data, Sample Selection, and Supply of Business Politicians

We provide details of sample construction, which involves the processing of tens of millions of web searches, in Appendix A. Here, we briefly summarize the main steps in the sample construction process and describe how the number of corporate executives seeking political office has changed over time.

2.1. Data on Business Politicians Based on Official Biographies

For each election cycle between 1980 and 2018, we obtain the official biographies of all members of the U.S. Congress (Representatives and Senators) and U.S. Presidents/Vice Presidents who were elected (or re-elected) in that election cycle (2,045 office holders in total). By reading their official biographies, we identify all federal office holders (i.e., members of Congress and U.S. Presidents/Vice Presidents) who, prior to being elected, held at least one position as the CEO (i.e., chief executive officer), president, chairman/chairwoman, or founder/owner of any private or public for-profit nonagricultural firm. We refer to such federal office holders as “business politicians,” of which there are 390. We also identify firms in which they worked prior to being elected. Table 1, Panel A, reports aggregate summary statistics on the share of business politicians among all federal office holders. Over the 20 election cycles between 1980 and 2018, this share stands, on average, at 15.8% and reaches its maximum of 22.6% in 2018 (Figure 1). On average, 67.4% of business politicians run as Republicans.

Table 1. Summary Statistics

We complement our sample of business politicians with data from a variety of other sources. For all members of Congress elected (or re-elected) between 1980 and 2016, we obtain data on their voting records and the bills that they introduce or cosponsor.⁵ We also collect campaign finance data for all candidates running for federal office during our sample period, regardless of whether they are ultimately elected or not.

2.2. Data on BoardEx Executives Running for Political Office

Although the sample of business politicians constructed from official congressional biographies enables us to study broad trends in executives’ political participation, it cannot inform us about the differences between business politicians and other politicians at the electoral stage (i.e., prior to becoming members of Congress); neither can it shed light on the extent to which the entry of executives into politics is supply- or demand-driven. To address these issues, we use the sample of corporate executives covered by BoardEx and track the incidence of people from this sample running for political office over time. BoardEx contains detailed biographies of all board members of S&P 1500 firms and provides employment details about all firms where a given individual worked, even if those firms are not part of S&P 1500.

The sample construction at this stage involves an algorithm that enables us to process tens of millions of search results. The details of this algorithm are provided in Appendix A; here we briefly describe the main steps. We start by identifying all people in BoardEx who have held, at any point in their BoardEx employment history, at least one position as the CEO, president, chairman/chairwoman, or founder/owner of any private or public for-profit nonagricultural firm (i.e., we choose the same types of executive positions that we use in constructing the sample of business politicians from official congressional biographies). This procedure leaves us with 61,502 unique individuals. We then build a web crawler to identify all instances in which any of these 61,502 executives run for political office.⁶ Having downloaded the relevant search results, we process them for keywords and then manually analyze the web pages that contain any mentions of corporate executives from BoardEx running for federal political office. In total, we perform 11,972,754 web searches and analyze 29,908,149 individual search results. Ultimately, we identify 167 campaigns for federal office between 1980 and 2016 in which BoardEx executives run. They win 98 of those 167 campaigns. The aggregate summary statistics for the sample of business politicians based on BoardEx are reported in Panel B of Table 1.

2.3. Supply of Business Politicians

Figure 1, where the solid red line depicts the share of business politicians among federal office holders, shows an increase in the number of such politicians after 2000. On average, there are approximately 72 business politicians holding federal office in any given election cycle between 1980 and 2000, and this average rises to 100 per election cycle in the period from 2002 to 2018. In fact, the share of business politicians among federal office holders has gone up in every election after 2000. Figure 1 also shows that the cost of congressional campaigns has risen during our sample period (the dashed blue line).⁷ This increase may have created a fundraising advantage for business politicians because they tend to be wealthier than the average citizen and because their rolodex is likely to include wealthy donors. Later in the paper we show that business politicians indeed enjoy a fundraising advantage over their opponents because they are more likely to self-fund their campaigns and because they benefit from contributions made by their firms.

Because the total number of federal office holders is fixed at 537 (435 U.S. Representatives, 100 U.S. Senators, one U.S. President, and one Vice President), the increase in the share of business politicians may have been accompanied by a decrease in the share of politicians with other types of experience. In Online Appendix B, therefore, we analyze three types of nonbusiness experience: service in the military, background in law, and agriculture experience. Although the share of politicians with experience in agriculture remained steady during our sample period, the share of politicians with a background in law and military service decreased (see Figure B.1), suggesting that business politicians replaced politicians with these types of experience.

Although most corporate executives (67.4%) run as Republicans, there are business politicians also among Democrats in every election cycle during our sample period. We explore the party affiliation of business politicians and its evolution over time in Figure 2. The bars in the figure show the share of business politicians among elected federal office holders separately for Republicans (the solid red bars) and Democrats (the shaded blue bars). The share of business politicians among Republicans averages 21.4% in our sample, whereas the share of business politicians among Democrats is 10.2% on average. The lines in the figure depict the share of Republicans among business politicians (the dashed green line) as well as the overall share of Republicans among elected federal office holders (the solid black line). As the figure shows, the recent increase in the number of business politicians cannot be solely attributed to the electoral strength of the Republican party because the share of Republicans in Congress remained fairly stable during our sample period. The share of business politicians, on the other hand, increased both among Republicans and Democrats, with most of the increase happening after 2000. In 1980, the share of business politicians among Republicans was 16.4% compared with 10.8% for Democrats. By 2018, these shares reached 31.7% and 14.4%, respectively.

Similar to the increase in the total number of business politicians elected to federal office, we observe an upward shift in the number of BoardEx executives running for office after 2000 (Figure 3). Specifically, the average number of BoardEx executives running for federal office in any given election cycle is about 7 in the period from 1980 to 2000, and this number rises to 11 per election cycle in the period from 2002 to 2016. This difference is statistically significant, suggesting that BoardEx executives are indeed more likely to run for office after 2000.

Is the increase in the share of business politicians supply- or demand-driven? On the demand side, it could be that the preferences of U.S. voters have shifted toward politicians with skills that executives are more likely to possess, such as business acumen, leadership, or the ability to run complex organizations.⁸ On the supply side, executives may be more likely to seek political office because the rising cost of political campaigns has given them a relative advantage in fulfilling their personal political ambitions or because the benefits from political participation for their firms have increased. In reality, supply and demand are likely to be closely linked. For example, voters may perceive business politicians as being of higher quality than other candidates (thus increasing the demand for business politicians), and this perception may in turn encourage more executives to run for office (thus increasing supply).

We do not attempt to separate supply from demand, both of which may have contributed to the aggregate trends we observe. However, we find some suggestive evidence that supply-side factors (i.e., a higher propensity of business executives to enter politics) have been more important than demand-side factors in explaining the recent rise in the number of business politicians. In particular, corporate executives’ likelihood of winning political office, while remaining high, has decreased over time. This pattern is the opposite of what one would expect had the increase in the number of business politicians been driven entirely by greater demand. Specifically, the likelihood that a BoardEx executive running for political office wins an election averages 77.0% between 1980 and 2000 but then falls to an average of 49.5% between 2002 and 2016. Furthermore, the average number of nonbusiness politicians running for office between 1980 and 2000 is neither economically nor statistically different from the average number of nonbusiness politicians running for office between 2002 and 2016, implying that the increase in the number of executives running for office cannot be explained by a higher overall propensity of ordinary citizens to enter politics.

3. Political Contributions and Business Politicians’ Early Fundraising Advantage

The increase in the number of business politicians running for federal office has coincided with an increase in the cost of political campaigns. Because executives are typically wealthy and have access to a network of wealthy donors, they may be in a better position, relative to nonbusiness politicians, to afford these rising costs. It is also likely that executives’ ability to self-finance their campaigns has increased over time because of a well-documented increase in executive compensation (Edmans et al. 2017).

In this section, we examine whether business politicians have a fundraising advantage over their opponents and the extent to which this advantage comes early in the campaign, which prior literature has shown may be important for electoral success (Bonica 2017).⁹ Our goal is to compare business and nonbusiness politicians running for office regardless of whether they ultimately win elections.¹⁰ Therefore, in this set of tests, we focus on the BoardEx sample because it consistently tracks a set of executives who run for office, regardless of whether they win elections.

Federal election law puts strict limits on campaign contributions. However, these limits do not apply to the funds a candidate donates or lends to his or her own campaign, implying that self-funding by wealthy candidates can potentially create a fundraising advantage. To examine whether there is such an advantage, we compare the extent of self-funding among business politicians and their opponents. We define self-funded candidates as those candidates that give at least $1 million in personal funds to their own campaigns.¹¹ As Panel A of Table 2 shows, self-funding is more prevalent among business politicians than among their opponents. In the sample of all candidates running in a given election, the share of self-funded business politicians is almost seven times higher than the share of self-funded nonbusiness politicians (23.6% compared with 3.5%). When we remove nonviable candidates by restricting the sample to the winners of primary elections, the differences in self-funding among business and nonbusiness politicians continue to persist (21.9% compared with 5.5%).

Table 2. Fundraising by Business Politicians and Their Opponents

Another source of business politicians’ fundraising advantage may come from their professional networks. Indeed, we find that business politicians raise more money from outside sources (i.e., sources that exclude self-funding) than do their opponents. In the sample of all candidates, business politicians receive $1.9 million in contributions from political action committees (PACs) and individuals compared with $848 thousand for nonbusiness politicians. In the sample of primary election winners, however, the total amounts business politicians and their opponents raise from outside sources are virtually identical. It is worth noting that these total amounts may conceal differences between business politicians and their opponents in terms of timing of their campaign fundraising, which is what we turn our attention to next.

To identify the dates when BoardEx executives and their opponents start their political campaigns, we rely on FEC Form 2 (Statement of Candidacy). This form, which has to be manually collected from the FEC website for each candidate, contains basic information about the candidate and the election in which he/she is running. When several Statements of Candidacy are available for the same candidate in the same election (e.g., an amendment is filed), we pick the form with the earliest filing date.

We compute the duration of an electoral campaign as the number of days between the filing date on the Statement of Candidacy and the election day. Based on this duration, we split each campaign into three equal time intervals (or stages): the beginning, middle, and end of campaign. We then collect contributions from PACs and individuals received by business politicians and their opponents during each stage of the campaign. Figure 4 shows that business politicians enjoy an early fundraising advantage over their opponents, both in the sample of all candidates as well as in the subsample of primary election winners. Figure 4 also shows business politicians’ cumulative fundraising advantage at each stage of the campaign, which we compute as the ratio of business politicians’ cumulative campaign contributions to their opponents’ cumulative campaign contributions. In the sample of all candidates, business politicians cumulative fundraising advantage persists throughout the entire campaign. The winners of primary elections, however, eventually catch up with their business politician opponents by the end of the campaign.

In Panel B of Table 2 and in Table 3, we examine business politicians’ early fundraising advantage more closely by analyzing campaign contributions received in the first 90 days after a candidate initiates his/her campaign. In the sample of primary election winners, business politicians enjoy an almost 2:1 fundraising advantage over their opponents in the first 90 days, and this advantage is even bigger in the sample of all candidates. The differences between business politicians and their opponents persist also when we analyze fundraising in a regression framework (in Table 3), which enables us to control for election cycle fixed effects and candidate characteristics such as incumbency and party affiliation.

Table 3. Regressions of Campaign Contributions Received in the First 90 Days on the Business Politician Indicator and Controls

The ability of business politicians to use their personal networks for fundraising may come from connections with their former firms. To investigate this possibility, we analyze campaign contributions that firms make to their executives running for office. We start with the sample of firms whose former or current executives run for political office at least once during our sample period and collect all campaign contributions made by these firms’ PACs and employees (we exclude candidate self-contributions from this analysis).¹² We collect these contributions for all election cycles in our sample, including the cycles in which none of the firm’s executives run for political office.

We find that firms become more politically active (in terms of campaign contributions) when their executives run for office. As Table 4 shows, the total amount of campaign contributions that a firm makes to all political candidates more than doubles (from $41,000 to $85,000) when one of its executives runs for office. This increase in firms’ total campaign contributions can be attributed almost entirely to the contributions they make to their executives’ electoral campaigns. In fact, in 77.6% of the cases when a firm’s executive runs for office, this executive is the firm’s highest supported candidate (i.e., the political candidate receiving the largest dollar amount of total contributions from the firm).

Table 4. Campaign Contributions Made by Firms Whose Executives Run for Political Office

We also find that firms provide larger financial support to their executives than to other politicians. On average, business politicians receive $53,000 from their firms (Table 5).¹³ In contrast, when none of the firm’s executives run for office, the average amount received by the firm’s highest supported candidate is $9 thousand.

Table 5. Firms’ Contributions to Business Politicians and to Highest Supported Candidates

In Table 6, we analyze corporate contributions in a regression framework, which enables us to include firm characteristics and a variety of fixed effects. The unit of analysis in these regressions is firm-candidate-election cycle: that is, we include all candidates that a given firm contributes to in a given election cycle, regardless of whether the candidate is the firm’s executive. The main explanatory variable in these regressions is an indicator set to one if the political candidate is the firm’s executive. This variable is set to zero if the candidate is not a business politician or is a business politician who is a former or current executive of a different firm.

Table 6. Regressions of Campaign Contributions on the Business Politician Indicator and Firm Characteristics

Regression results show that firms support their executives’ political campaigns even after we include election cycle, firm, and politician fixed effects. The estimates from our most stringent specification that includes the full set of fixed effects and control variables imply that executives receive, on average, $27,000 more from their firms compared with other politicians. This additional amount is more than five times larger than the unconditional sample average. These results are also robust to using a log transformation of campaign contributions (see Online Appendix B).

The financial support that firms provide to their executives’ electoral campaigns is consistent with the idea that these firms expect to benefit from their executives’ political wins. However, there may be other reasons for firms to support their former executives’ electoral campaigns. Such campaigns, for example, may serve as a cheap way to publicize the firm.¹⁴ To investigate this possibility, in Online Appendix B, we examine Google Trends that mention the executive’s name in conjunction with his/her firm when this executive announces his/her run for office. Although we do observe a spike in Google searches that mention the executive’s name alone or in conjunction with the name of the firm, this spike is rather transient and disappears almost entirely by the second week after the announcement. Furthermore, there is virtually no observable impact on Google searches for the firm’s name alone, suggesting that executives’ political campaigns generate little additional publicity for their firms. We conclude that this channel alone cannot fully explain firms’ willingness to contribute to their executives’ political campaigns and that there are other benefits that accrue to firms when their executives win political office. We therefore turn our attention to firm-value implications of business politicians’ electoral wins.

4. Firm-Value Implications

To assess firm-value benefits generated by business politicians, we compute cumulative abnormal stock returns (CARs) around the dates on which business politicians win political office. We also investigate if there are any spillovers from business politicians’ electoral wins to other firms in the same industry.

Table 7 reports the CARs for firms in which business politicians worked prior to running for office.¹⁵ The average CARs are positive, indicating that the market participants expect firms to benefit from having their former executives in elected office. The returns in our baseline model (reported in Panel A) range from 0.9% to 1.8%. In terms of market capitalization, an average firm in our sample gains approximately $200 million within three days after its executive wins an election.¹⁶ In Panel B of Table 7 and in Online Appendix B, we show that these results are robust to winsorization and are similar if we use the Fama-French three-factor model instead of the market model.¹⁷ In Online Appendix B, we also provide some evidence that firms whose executives hold large equity stakes experience higher equity returns around elections than firms whose executives hold small equity stakes. This evidence, although indirect, suggests that equity ownership may align the interests of business politicians with the interests of their firms.

Table 7. Firm-Value Implications of Business Politicians’ Winning Political Office

Next, we analyze whether the election of business politicians generates spillovers to other firms in the same industry. For each firm where a business politician worked prior to being elected, we identify up to 10 firms in the same four-digit Standard Industrial Classification (SIC) code that are closest to the business politician’s firm in terms of market capitalization.¹⁸ We then compute these firms’ CARs around the elections when business politicians win office. Our evidence (reported in Panel C and Panel D of Table 7) points to positive industry spillovers: that is, when a business politician is elected, his/her firms’ industry peers experience positive abnormal stock returns, which range from 0.6% to 0.9% in our baseline specification.

One mechanism through which business politicians, once elected, may generate value for their firms is their service on congressional committees. These committees have wide discretion in determining which bills ultimately become law and perform an important oversight function (Shepsle and Weingast 1987). Committee members, therefore, can have a substantial impact on firm value. For example, stock prices of major health insurance companies dropped by as much as 5% in April 2019 in response to a mere suggestion by Chuck Grassley, the then chair of the Senate Finance Committee, that the committee was considering new laws to regulate the market for prescription drugs.¹⁹ Although business politicians’ ability to join their preferred committees is not unlimited, existing congressional procedures give individual members considerable freedom in seeking committee assignments, subject to the constraints of seniority (Schneider 2006).²⁰ Our results, reported in Table 8, suggest that the probability of a business politician joining a congressional committee more than doubles if this committee has jurisdiction over the industries where the business politician worked prior to being elected.²¹

Table 8. Business Politicians’ Service on Congressional Committees

Apart from service on congressional committees, business politicians may also affect firm value through legislation, the importance of which has been recognized in the literature (Cohen et al. 2013, Neretina 2019). To explore this possibility, in Online Appendix B, we analyze the CARs around the dates when legislation introduced by business politicians is passed in Congress. Following Neretina (2019), we identify the dates when bills clear the second chamber but are not yet signed into law by the President. Consistent with the idea that business politicians propose legislation favorable to their firms, we find that bill passage is associated with positive stock returns, which range from 1.1% to 2.3%.

Overall, our results indicate that corporate executives’ entry into politics benefits their firms and industries. In addition to firm- and industry-specific effects, however, business politicians may also have an aggregate impact on the U.S. policy and business climate more generally. It is this aggregate impact that we investigate next.

5. Aggregate Impact of Business Politicians on the Legislative Process

Because most politicians (including business politicians) introduce few bills, perhaps the biggest impact they can have on aggregate policy is by voting on bills introduced by others. Our next set of tests, therefore, aims to examine business politicians’ overall voting records.

To quantify politicians’ voting records, we use data on interest group ratings. These ratings provide a numerical score for each politician indicating how closely the politician’s voting record is aligned with the interests of a particular interest group. Probusiness ratings are provided by the Chamber of Commerce of the United States (CCUS); prolabor ratings are provided by the Committee on Political Education of the AFL-CIO (COPE); and proconsumer ratings are from the Consumer Federation of America (CFA). We also collect the overall liberal/conservative scores (DW-NOMINATE) developed by Poole and Rosenthal (1991) and available from the Voteview database (Lewis et al. 2019). Probusiness, proconsumer, and prolabor ratings range from 0 to 100, with a higher score indicating a stronger alignment with the preferences of the given interest group. The original DW scores range from –1 to +1, with a larger (positive) number indicating a more conservative voting record. We multiply DW scores by 100 to make their scale comparable to the other scores.

We are interested in business politicians’ causal impact on legislation. Identifying this impact is challenging because only politicians who ultimately win office can pass legislation. Therefore, the observable voting records may reflect not only business politicians’ own views on legislation but also the aggregate preferences of their voters. It may be, for example, that business politicians are elected as a consequence of aggregate shifts in voter preferences toward the policies advocated by such politicians.²² It is therefore important to separate the preferences of the electorate from the preferences of politicians, especially in light of recent evidence in Matsusaka (2021), who shows that legislators often vote their own preferences when their personal views diverge from the majority of their constituents.

To separate the legislative impact of business politicians from the aggregate preferences of their constituencies, we perform two sets of tests. First, we compare the voting records of business politicians to the voting records of their predecessors and successors. Second, we analyze the voting records of business politicians who win close elections.

To compare business politicians to their successors and predecessors, in Figure 5, we plot the voting patterns within the same constituency before, during, and after a business politician is in office. The figure is constructed in normalized event time to account for the fact that different business politicians serve a different number of terms. Time 0 indicates the first year that a business politician is in office, time 2 indicates the last year that a business politician is in office, and time 1 represents the average of a business politician’s voting score across all years during which he/she is in office. This procedure enables us to consistently compare voting scores of different politicians regardless of how many terms they serve. Time –2 and time –1 show the average voting scores of the business politician’s predecessor, whereas time 3 and time 4 show the average voting scores of the business politician’s successor, all within the same constituency. The figure includes four panels: the top two panels contain liberal voting scores associated with proconsumer and prolabor interest groups, whereas the bottom panels contain conservative voting scores associated with probusiness interest groups and the overall conservativeness of a politician.

In all four panels in Figure 5, we observe sharp changes in voting scores immediately after a business politician assumes office and immediately after he or she leaves office, which suggests that the voting behavior of business politicians cannot be fully explained by permanent long-run political preferences of their constituencies. Relative to their immediate predecessors and successors, business politicians consistently vote in line with corporate interests and against proconsumer and prolabor interests.²³ This evidence suggests that, in addition to generating firm- and industry-specific benefits discussed earlier, business politicians also shape aggregate policies in a way that favors corporate interests.

The graphical evidence in Figure 5 suggests that permanent long-run political preferences of voters cannot fully explain the behavior of business politicians. However, this evidence is not sufficient to rule out the influence of temporary shifts in voter preferences on the behavior of business politicians. Nor does this evidence allow us to isolate the effect of party affiliation. We address these challenges in our next set of tests, in which we compare the voting scores of business and nonbusiness politicians who win close elections (we use the victory margin of 10% to identify close elections).

To the extent that the outcomes of close elections are determined by chance, the policies that the winners of such elections implement cannot be attributed solely to the shifts in aggregate voter preferences. After all, had all voters supported business politicians’ policies, the elections of such politicians would not have been close. Therefore, the differences in voting scores between business and nonbusiness politicians who win close elections should identify the causal effect of business politicians on legislation, free from the contaminating influence of aggregate voter preferences.

In Tables 9 and 10, we regress the voting scores of winners of close elections on the business politician indicator variable as well as party affiliation (i.e., the Republican indicator variable) and election cycle fixed effects.²⁴ Consistent with the graphical evidence presented earlier, we find that business politicians are more likely than nonbusiness politicians to vote against the interests of labor unions and consumers. Business politicians also accumulate a more conservative overall voting record relative to their nonbusiness peers, even after controlling for party affiliation. The effect of business politicians on probusiness policies, although economically sizeable, is not statistically significant in the sample of close elections. In some specifications in Tables 9 and 10, we directly control for aggregate voter preferences by including the Republican vote share as an explanatory variable. This variable never enters significantly, suggesting that our identification strategy is successful at removing the effects of aggregate voter preferences. In Online Appendix B, we show that the results reported in Tables 9 and 10 continue to hold also when we drop cases in which both candidates in the general election are business politicians. We remove such cases to avoid instances in which the election winner is bound to be a business politician regardless of the vote outcome.

Table 9. Impact of Business Politicians on U.S. Legislation: Evidence from Close Elections, Liberal Ratings

Table 10. Impact of Business Politicians on U.S. Legislation: Evidence from Close Elections, Conservative Ratings

In addition to the evidence from close elections, we perform similar regressions on the full sample and report the results in Tables 11 and 12. Although the results in Tables 11 and 12 are less well identified than those reported in Tables 9 and 10, the two sets of results are not dissimilar. In the full sample, we find that the propensity of business politicians to favor corporate interests is not only economically but also statistically significant.

Table 11. Impact of Business Politicians on U.S. Legislation: Large-Sample Evidence, Liberal Ratings Evidence

Table 12. Impact of Business Politicians on U.S. Legislation: Large-Sample Evidence, Conservative Ratings

Overall, our evidence suggests that business politicians shift the balance of power toward corporate interests by supporting probusiness legislation and voting against the legislation supported by labor unions. Importantly, this policy shift cannot be attributed entirely to the changes in the underlying voter preferences.

6. Conclusion

We document a rising trend in the number of corporate executives seeking political office in the United States and find evidence that the election of business politicians affects their firms, industries, and aggregate U.S. policy.

At the electoral stage, business politicians enjoy a fundraising advantage over their opponents, especially early on in the campaign. Part of this advantage can be attributed to a greater propensity of business politicians to self-fund their campaigns and to campaign contributions from their firms, which donate six times more to their executives running for office than to other candidates.

We also find that the electoral wins of business politicians are associated with positive stock returns for their firms and industries. Once elected, business politicians are likely to join congressional committees with jurisdiction over their industries.

Finally, we examine business politicians’ aggregate voting records and find that they generally vote in favor of business interests and against the interests of labor unions. Using close elections for identification, we show that business politicians may change aggregate policies, even if voters do not necessarily demand those changes. Overall, the entry of corporate executives into politics appears to have shifted the balance of power toward corporate interests.