Social ties and the political participation of firms

Abstract

It is widely accepted that better-connected firms exercise more political influence. However, because both social ties and political influence are difficult to observe, it remains little understood how different types of ties affect different types of political action. We develop new theory that contrasts the effects of ties to other firms in the same industry (peer ties) with ties to elected officials and bureaucrats (government ties). We argue that peer ties are useful for facilitating collective action, which is most often used to seek influence over broad policy issues that affect many firms. In contrast, while government ties are more versatile, these should primarily also allow firms to pursue narrow, particularistic issues that benefit only one or a small number of firms. Using a new survey of foreign-owned firms operating in the Philippines, we demonstrate that peer ties play an important role in firms’ efforts to influence policy at the national level, where issues are more likely to affect large numbers of firms. Government ties are also valuable, though primarily at the local level, where issues are more likely to affect a smaller number of firms.

1 Introduction

The social ties and connections of firms shape a variety of political and economic outcomes.¹ This is especially important in countries with weakly institutionalized political systems, where informal institutions, including social networks, are often used to substitute for weak formal institutions.² The existing literature on social networks and firms’ political participation has focused almost exclusively on ties between firms and government officials (see, e.g., Fisman 2001; Faccio 2006; Malesky and Taussig 2009). But firms operate in a much richer social landscape beyond their ties to political actors: firms also have clients, suppliers, consultants, and peer firms in the same industry. In particular, we show that peer ties, which connect owners, managers or board members at one firm to counterparts at other firms in the same industry, are important determinants of firms’ political participation.

Our focus on peer ties breaks with the existing literature on firm collective action, which often focuses on industry size or industry concentration (e.g. Olson 1965; Grier et al. 1994; Hansen et al. 2005). While firms in the same industry often compete with each other, they also tend to share many policy preferences, and firms often engage in political action at the industry-level. Peer ties are interpersonal relationships between owners or managers at one firm and owners or managers at another firm in the same industry; thus, the strength of peer ties varies from firm to firm, not just from industry to industry. Peer ties may form in a number of ways: through family relationships, longstanding friendships, or formed through repeated interaction in trade associations or other professional venues. One particularly common linkage between managers and owners at different firms is through their alumni and professional networks.

Firms in the same industry often compete with one another for market share and they vary across import/export behavior and other dimensions such that their policy preferences do not always align (e.g., Baccini et al. 2017). However, we expect that political cooperation between peer firms is valuable because we expect firms in the same industry, whether or not they are direct competitors in the same local market, often share collective interests in public policies that benefit the whole industry. Social ties between peer firms enable social sanctions and rewards, which can be used as selective incentives to facilitate collective action. For example, firms may collude in refusing to do business with or undermining the reputation of a firm that refuses to participate in a collective lobbying effort. Thus, we expect ties to peer firms to be associated with a greater likelihood that firms to act collectively in pursuit of broad policy objectives that benefit large groups of firms.

Government ties are direct links between firms and government officials, such as when current or former government officials own, manage, or sit on the board of directors of a firm. Government ties may be useful for pursuing many types of political influence, and we expect these ties may provide firms with information as well as opportunities to seek influence directly. However, we expect that direct government ties have their largest effect on firm behavior when firms use them to seek narrow or particularistic policy outcomes that benefit one or a handful of firms, such as securing a permit, a regulatory exemption, or a local infrastructure investment. Given the costs of calling in favors via government ties, we expect firms to prefer to draw on these ties in pursuit of narrowly concentrated benefits that do not also benefit their competitors.

Because of the focus on government ties and the pursuit of particularistic benefits, the existing literature on social ties and firms’ political influence has been closely connected to the study of corruption (e.g. Fisman 2001; Malesky and Taussig 2009; Dinç 2005; Sojli and Tham 2017). By extending our analysis to peer ties, we show that social ties are also central to the ability of firms to act collectively to seek influence over large-scale national policies. Our theory implies that the density of peer ties between firms is a crucial determinant of the influence that business interests yield in national level political debates over tax rates, trade policy, climate change, labor protections, and other foundational policy issues. To understand the balance of influence between business interests, labor interests, and other groups in society, we must understand how social ties shape the ability of firms to act collectively in pursuit of their shared interests.

A lack of data helps explain why most past scholarship has failed to evaluate the role of peer ties between firms. Firms’ peer ties are extremely difficult to measure. While a recent groundbreaking paper examines the effect of supply-chain relationships on the probability that a foreign firm engages in a formal dispute with the host government (Johns and Wellhausen 2016), we are unaware of any existing empirical work evaluating the effects of firms’ peer ties on political participation. Indeed, when firms’ propensity to act collectively to seek political influence has been evaluated, these evaluations have used industry concentration, rather than social ties, as the key causal variable (e.g. Hansen et al. 2005).

To measure firms’ social ties, we use new data from face-to-face interviews with the managers of 237 foreign-owned firms in the Philippines. These survey data allow us to assess firms’ ties to government officials and peer firms, and to measure peer ties in two ways – both number of ties and frequency of interaction. To assess peer ties, we ask respondents for the number of peer firms in their industry with which their firm has an ongoing business relationship.³ We also ask them about the frequency of their firm’s interaction with other firms in their sector. To measure government ties, we ask whether any of the firms’ owners, managers, or members of the board of directors have held a range of different government posts, from bureaucratic posts and local municipal offices up to governorships and senate seats. We also use the same frequency of interaction question with respect to elected officials in local and national government and members of the bureaucracy.⁴

Measuring firms’ political participation is also extremely difficult. Thus, existing studies assess either easily observable forms of political participation, like campaign contributions, or they capture the effects of political influence, like firm stock price (Fisman 2001), government bailouts (Faccio 2006), or access to credit (Malesky and Taussig 2009). Because we examine self-reported political participation, we can pick up informal means of engaging with the government in a way that was not possible in previous work. Just as business can be conducted on the golf course or agreements sealed with handshakes rather than signatures, influence can be pursued without campaign donations or other formal actions. We expect that the decision to measure political participation via campaign contributions has lead the field to systematically under-estimate the extent of firms’ political participation and, in particular, to underestimate the importance of the types of informal influence that social ties help facilitate.

While our new survey-based measures of political participation offer a significant improvement over existing measures based on lobbying and campaign expenditures, they also have their own limitations. Our theory makes predictions regarding the types of firms most likely to pursue narrow, particularistic policy goals versus broad policy goals that benefit large numbers of firms. However, specific policy objectives are too sensitive to ask firms about directly, given public relations risks and concerns about corporate espionage. Instead, as a proxy, we ask about whether firms sought political influence at the national and/or the local level. To test our theory, we rely on the assumption that national-level policy change is more likely to affect large numbers of firms, while local-level policy changes are likely to affect smaller numbers of firms.

This assumption rests on the specific nature of political decentralization in the Philippines. In the Philippines, the national government controls policy levers that generally affect large groups of firms, like tax rates and labor and environmental regulations (Llanto 2012; Capuno 2012). Local governments rely on central government transfers for most funding, but have authority over localized infrastructure spending, distribution of development funds, allocation of many types of permits, and the actual enforcement of many business regulations (Llanto 2012; Capuno 2012). Thus, firms are best able to achieve firm-specific perks, such as tailored development funding and regulatory exemptions at the local level.

Of course, even in the Philippines, it is possible that firms sometimes seek particularistic benefits at the national level and broad collective benefits at the local level. To the extent that this occurs, it should weaken our ability to differentiate between the effects of peer ties and government ties. Thus, the fact that we find the distinct pattern of relationships we predict between types of social ties and types of political participation provides strong evidence for the underlying theory.

A common difficulty in studies of social and political networks is the fact that these network ties are not randomly allocated. In our case, firms freely choose to invest in different types of relationships and these choices may be affected by our variables of interest. While an experiment assigning network ties to firms is not feasible in our context, we can partially address these concerns by controlling for factors that can affect these incentives, like firm size and profitability. We also control for sector with a vector of indicator variables that absorb the effects of industry-level factors that affect the need for political influence, like the extent of regulation in a given industry.

Our empirical approach also provides an avenue for (partially) addressing the difficulties associated with reverse causation–i.e., the possibility that firms perceived to be more politically active may be more attractive to other firms seeking connections. Our research design contrasts the effects of ties to peer firms with the effects of ties to political actors, allowing us to show a pattern of correlations that is not readily explained by reverse causality. If reverse causation were driving our results, we would expect to find that peer ties were correlated with political participation at both levels, not the national level alone.

We also conduct additional sensitivity tests to support our results. First, we conduct a placebo test in which we show that a type of social tie that we would not expect to facilitate political efforts – i.e. client ties – are indeed not significantly associated with political participation. This lessens concerns that the empirical patterns we observe are simply driven by more resources going to socially connected firms. Second, we conduct a falsification test to show that peer ties and government ties do not affect attempts to influence levels of government among firms with no incentive to do so. To do this, we exploit the fact that special economic zones (SEZs) reduce incentives for firms to engage with local government (local government has limited ability to tax and regulate these firms) and show that firms in SEZs use their social ties to engage with government at the national level only.

Our analysis represents an important step forward in an area where data are scarce and difficult to collect, but there remain limitations on the strength of inference our results allow. We regard the evidence we present here as only a first step – consistent with our theory and suggestive of the importance of different types of firm ties for understanding political influence, but leaving room for future work evaluating the effects of a wider range of firm ties on political strategies.

The next section presents our theoretical framework. We then introduce our research design, data, and present empirical results. Finally, we discuss the implications of our findings and directions for future research.

2 Theory and literature

Social ties are critical to firm performance in both politics and markets. Lack of necessary social ties can be a major impediment to foreign firms operating in emerging markets in particular (e.g. Johanson and Vahlne 2009, Graham 2019). Prior work on firms’ political participation has focused exclusively on ties to government officials (e.g. Fisman 2001; Faccio 2006; Hillman 2005). Firms with government ties secure more funding from state-owned banks (e.g. Malesky and Taussig 2009; Dinç 2005) and more government contracts (e.g. Sojli and Tham 2017). The revolving door literature has shown that ties to the bureaucracy can also benefit firms, including via more favorable regulatory treatment (e.g. Gormley 1979; Luechinger and Moser 2014). In contrast, the literature on peer ties between firms has ignored political participation and focused instead on firms’ ability to use peer ties to manage peer relationships. Most notably, firms can use peer ties to better screen business counterparts (e.g., McMillan 1997) and enforce contracts (e.g., Greif 2006, Hadfield and Weingast 2014).

We develop new theory specifying the role of peer ties in facilitating political participation by firms. We then contrast the role of peer ties with the role of government ties, which we expect serve a different, though also critical, purpose. We argue that peer ties are primarily useful in facilitating firms’ ability to act collectively in pursuit of shared policy goals. These goals may include seeking trade protections (e.g., Goldberg and Maggi 1999), limiting regulation or weakening enforcement (e.g., de Figueiredo and Tiller 2001), reducing corporate taxation (e.g., Shin 2017), increasing wage protection for workers (e.g., Watson and Arunachalam 2018), or other policy changes that benefit large groups of firms, such as an entire industry.

The link we draw between peer ties and firms’ ability to act collectively in pursuit of broad policy goals facilitates a long-overdue revision of Olson’s (1965) seminal theoretical work on firms’ political participation. In Olson’s formulation, concentrated industries with fewer firms have an easier time engaging in collective political participation because smaller groups can better use selective incentives to deter free-riding. Unfortunately, decades of empirical work has failed to find support for this expectation – this failure is known as the paradox of corporate political participation (Hansen et al. 2005). In our theory, firms use peer ties to facilitate the use of social sanctions and social rewards as selective incentives for the same purpose. In other words, even if Olson is correct that the capacity for collective action is crucial to firms’ national level political participation, we show that concentrated industries are not the only mechanism for facilitating collective action. Our work highlights the peer ties among firms as a crucial and overlooked driver of firms’ ability to act collectively in pursuit of national-level policy objectives.

While peer ties are excellent for facilitating collective action in pursuit of policies that benefit large numbers of firms, we expect them to be less useful for pursuing narrow, particularistic benefits. It is difficult to act collectively in pursuit of ends that do not benefit most members of the group. However, we do expect that government ties are quite useful for seeking these narrow benefits.

Asking for favors from members of one’s social network is costly as favors incur obligations. Thus, it is most efficient to use government ties to seek particularistic benefits. If a former congressperson or mayor sits on a corporate board, for example, that direct government tie provides the firm with disproportionate access to government officials and possibly influence over their actions. But calling in favors from former colleagues imposes costs on that board member. Thus, we expect that firms wish to concentrate the benefits of any policy influence they achieve through government ties, reaping the benefits themselves rather than diluting them and possibly benefitting competitors.

For example, we expect firms with government ties to pursue approval for a specific permit, zoning variance, or regulatory exemption. While government ties are no doubt useful in pursuing broad policy changes as well, we expect firms to be reluctant to deplete this scarce resource in pursuit of policy changes with diffuse benefits, like decreased regulation for an entire industry. Consistent with this logic, most of the extant literature on firms’ use of social ties to influence policy involves firms’ use of direct ties to government officials to secure private benefits, such as loans from state-owned banks (e.g. Dinç 2005; Malesky and Taussig 2009).

2.1 Local vs. National Government

One reasonable proxy for the scope of the issue over which firms seek influence is the level of government whose actions they seek to affect. This is especially the case in decentralized contexts such as the Philippines. National-level policies are generally broad in scope, affecting large numbers of firms, while local-level policies and regulatory choices usually affect fewer firms. National-level policy governs nationwide tax rates, environmental regulations, labor regulations, and other policies with broad effects. By contrast, local governments tend to focus on narrower policies that are more compatible with firm-specific benefits: issuance of permits, regulatory forbearance, and similar policies that can be targeted to individual firms.

In Federalist 51, James Madison contrasts the ability of narrow factions to dominate within a state to the inability of any single faction to do so at the national level. At the national level, “among the great variety of interests, parties, and sects which it embraces, a coalition of a majority of the whole society could seldom take place on any other principles than those of justice and the general good.” Thus, when firms seek policy change at the national level, they generally must pursue that change as part of large coalitions, and thus must seek change that benefits a broad swath of firms. Conversely, when firms seek particularistic policy changes tailored to the needs of their firm, they are less able to draw effectively on ties to peer firms. Instead, firms must rely on direct ties to political actors, including both bureaucrats and policymakers.

Because peer ties facilitate collective action, we expect them to be useful at the national level where policy effects are broad, but not the local level, where effects tend to be more narrow. In contrast, we expect government ties to be used primarily at the local level, where the effects are most concentrated, and be used more sparingly at the national level, where the benefits are more diffuse.

H1 Peer-National Hypothesis: Firms with more and stronger ties to peer firms are more likely to attempt to influence national-level policy in the investment host country.

H2a Government-Local Hypothesis:Firms with more and stronger ties to bureaucrats and elected officials are more likely to attempt to influence local-level policy in the investment host country.

H2b Peer-Local Hypothesis: Ties between peer firms have little or no effect on firms’ propensity to seek local-level policy influence in the investment host country.

We acknowledge that, even in decentralized contexts, the distinction between local- and national-level policy is an imperfect proxy for issues that affect individual firms versus large numbers of firms. However, to the extent that firms sometimes seek particularistic benefits at the national level, and to the extent that firms seek policy change that benefits large groups of firms at the local level, this serves as a signal-weakener in our analysis. Thus, if we find support for our hypotheses despite the imperfections of our measure, it suggests that the underlying mechanism is quite strong.

2.2 The paradox of corporate political participation

Perhaps the most widely held theoretical expectation in the existing literature on corporate political participation is Olson’s (1965) claim that firms in more concentrated industries are better positioned to collectively seek political influence. Surprisingly, however, there is little evidence in the U.S. that firms in highly concentrated industries are more politically active than those in less concentrated industries (e.g. Hansen et al. 2005). We argue that this paradox arises for two reasons. First, studies of industry concentration omit the crucial role that informal peer ties play in facilitating collective action among firms. Even if Olson is correct that collective action is crucial to firms’ national-level political influence, peer ties may be a more effective driver of collective action than industry concentration. Second, past studies focus on easily observable channels of political participation like campaign donations, while unable to account for less observable actions (including those that rely directly on social ties). By measuring only a small subset of political participation actions, existing scholarship may underestimate the effect of industry concentration, and collective action more broadly, on firms’ political participation.

Hansen et al. review 15 earlier tests of Olson’s proposition, all of which measure political activity as political action committee (PAC) formation, PAC contributions, and/or campaign contributions. Hansen et al. also examine a dummy variable for lobbying expenditures as a dependent variable. In all cases, the scope of the analysis remains restricted to some form of expenditures as a measure of attempts to influence policy. None of these analyses capture influence-seeking that proceeds through any channels except direct spending. There is a good pragmatic reason for scholars to take this approach. Data on many types of campaign, soft money, and political action committee (PAC) donations, as well as certain types of lobbying expenditures, must be publicly disclosed in the U.S., which make these data easy to obtain. However, financial contributions and expenditures on lobbyists are not the sole avenues, and perhaps not even the primary avenues, through which firms seek to influence policy, even in a developed democracy like the United States.

Well connected firms may seek influence through less-observable channels, most notably through “revolving doors” in which politicians and bureaucrats move from industry into policymaking positions and then back into industry (e.g. Luechinger and Moser 2014). Politicians may value the implicit promise of lucrative future employment far more than they value campaign contributions. With regard to lobbying expenditures, the most valuable lobbyists are typically individuals with close personal ties to politicians (Bertrand et al. 2014; Vidal et al. 2012). Firms with well-connected executives and board members may be able to lobby effectively without employing outside lobbyists; their executives and board members lobby directly. Thus, we expect that the relationship between lobbying expenditures and effort to influence policy is tenuous at best.

2.3 Addressing the possibility of reverse causation

Our central theoretical prediction (H1) is that peer ties facilitate collective action and thereby enable political influence-seeking at the national level. Our peer-local hypothesis (H2b) complements this expectation by helping us rule out the possibility that reverse causation might cause us to find support for H1 even if our theory is incorrect. H2b asserts that firms with more peer ties are no more likely than other firms to seek influence at the local level.

The risk of reverse causation stems from the possibility that firms seek ties to other firms perceived to be politically active, in expectation that ties to politically active firms may somehow provide second-hand benefits. However, if firms seek out ties to politically active firms, they should equally seek out ties to firms that are active at the national or the local level – if the logic of reverse causation holds at all, there is no reason to expect it to hold only for national-level efforts. Thus, if reverse causation is strong, it should induce a positive correlation between peer ties and political participation at both the national and the local level. This produces a useful empirical contrast. If our theory is correct, we should observe a positive correlation between peer ties and national-level political participation (H1) and no correlation between peer ties and local-level political efforts (H2B). However, if reverse causation is strong, we should see a positive correlation between peer ties and participation at both the national and local level. Thus, if we find that peer ties induce firms to seek policy influence only at the national level and not at the local level, it suggests that our results are not simply driven by reverse causation.⁵ In Table 3, and throughout our empirical results, we find exactly this pattern.

There is also a second potential type of reverse causation at play, but this is a type of reverse causation that is consistent with our theory and, indeed, can only exist if firms believe our theory to be correct. If our theory is correct, we would expect firms that intend to seek political influence to invest in developing the social ties necessary to do so effectively. However, if our theory is incorrect and social ties are not useful in pursuing political influence, then firms that intend to participate politically would have no reason to invest in social ties as a tool for doing so.

While we expect that firms’ strategic choices to invest in developing social ties drives some of the correlation we observe between peer ties and national-level participation, we do not expect it accounts for all or even most of it. Firm networks develop over long periods of time and depend on corresponding choices and actions of other firms, which means that they are difficult to change quickly and cannot be wholly determined by the firm leadership. Thus, while we do expect managers to develop ties strategically to facilitate political participation, this is a long-term process that cannot always respond quickly enough to match firms’ changing political needs.

To mitigate the risk posed by this type of reverse causation, we control for a wide range of factors that affect firms’ strategic incentives to form either peer or government ties. Most importantly, we include a vector of dummy variables for different industries. This removes the effect of industry-level factors that affect firms’ strategic need for social ties, such as how heavily regulated a particular industry is.

3 Empirical strategy

Our empirical analysis covers the characteristics and behavior of foreign firms doing business in the Philippines. A crucial advantage of working in the Philippines is that the division of authority between local and national government allows us to test our theory without querying firms about the specific policy issues over which they seek influence. The Philippines is also similar to other developing countries in that politics is characterized by clientelism (Hicken and Kuhonta 2014; Montinola 1999).⁶ Family firms and networks of political influence dominate the business landscape, and these networks are characterized by dense ties between firms and government (McCoy 1993; Abinales and Amoroso 2017). It is quite common for current or former government officials to sit on firm boards, or to serve as firm owners or managers, providing direct ties to government. Bribes and other informal incentives are also frequently used to facilitate business interactions (Mendoza et al. 2015). The practice is so pervasive that many firms will have a specific line item in their budget for “facilitation fees” or “representation fees” for these purposes (Author interviews, 2014). While a single-country study is limited in terms of external validity, findings from the Philippines are informative for understanding many other countries with similar challenges for promoting business growth.⁷

The Philippine 1991 Local Government Code outlined separate responsibilities for local and national government with implications for the business environment (Capuno 2010). At the national level, a number of agencies and departments cover large-scale tax policy and business regulation. In addition to the Departments of Finance, Trade and Industry, Labor and Employment, and Agriculture, many heads of sector- and region-specific economic development agencies hold cabinet level ranks.

At the local level, a number of business-relevant spending, service provision and regulatory enforcement responsibilities were devolved to municipalities, including the repair and maintenance of local infrastructure, issuance of permits and licenses, and provision of agricultural, fishery, mines, and geoscience services (Llanto 2012; Capuno 2012). While municipalities can also raise their own revenues through local taxes and fees, local governments rely primarily on transfers from the central government, which provide 85 percent of their budgets on average (Troland 2014). These funds are disbursed in tranches, and mayors have substantial discretion over how these funds are spent (Hutchcroft and Rocamora 2003; Abinales and Amoroso 2017).⁸ Thus, firm-specific benefits such as permits, tailored service and infrastructure spending, and regulatory forbearance are all best sought at the local level.

To provide one example, the mining industry in the Philippines is governed by strict, national-level environmental regulations. Most firms in the industry would benefit from collective political action to reduce the stringency of these regulations. However, many firms also address the issue via local level political participation, seeking regulatory forbearance at the municipal level. A study using remote sensing to identify mining sites found a number of illegal mining sites and processing facilities, apparently operating quite freely (Chaussard and Kerosky 2016). In some areas, mining firms direct so much money to local political races that citizens can receive as much as $125USD from mayoral candidates seeking their electoral support (Cruz et al. 2018; Author Interviews, 2013). The return to firms for their local-level spending is a willingness of local governments to selectively forgo enforcement of environmental regulations against the preferred firms, providing an enormous firm-specific advantage (Teehankee 1993).

While our theory applies to both foreign and domestic firms, our use of a dataset covering foreign firms exclusively means that our results are less likely to be confounded by firms with the hardest-to-observe multi-generational networks of political influence. While one might take for granted that social ties are extremely important to, and prevalent among, domestic firms in a context like the Philippines, scholars have long recognized that being outside of these networks is a key liability facing foreign firms (e.g. Johanson and Vahlne 2009). Our data show that, even among multinational firms, social ties are strong predictors of firms’ political participation.

3.1 New survey data

To address the scarcity of data on firms’ social ties and their efforts and success in influencing policy, we introduce new survey-based measures of both. These data are drawn from a 2014 survey of 237 foreign-owned firms operating in the Philippines. Respondents were all firm managers, and they spoke face-to-face with enumerators for roughly forty minutes, answering a variety of questions about firm characteristics, social connectedness, and firm strategy.

Collection of new data for this project was essential because existing firm-level data sources lack the measures of interest in this study, i.e. firms’ social connectedness and their efforts to influence policy. Most of the firm-level data employed in the existing literature comes from databases such as Compustat and Bureau Van Dijk that aggregate information on publicly traded firms. These databases lack any information on social ties and the information they provide on political strategy is limited to data on lobbying expenditures and campaign contributions.

By contrast, detailed surveys are rare, usually avoid sensitive topics, and are typically limited to small and medium-sized enterprises. In one survey that laudably managed to gather sensitive information, Mendoza et al. (2015) conducted surveys of small firms in the Philippines to assess the extent of corruption and its effects on firm performance. While small and medium-sized enterprises may be more accessible to enumerators conducting surveys, there may be less variation in terms of their political strategies.

Questions about firms’ social ties and their efforts to influence policy are both sensitive (Jensen et al. 2010), and our survey was extensively piloted to reduce respondent concerns about corporate espionage and public relations risks. As a result, while we were able to ask respondents about the level of government at which their firms had sought policy influence, we were not able to ask about the specific policy goals they pursued. Thus, while our new data enable analysis that greatly advances our understanding of firm political participation, a test using exact policy objectives will require even more detailed data.

Even when respondents are willing to answer sensitive questions, another concern is that social desirability bias may affect responses. In Section C of the Online Appendix, we report the results of a list experiment we ran with regard to the most sensitive question that remained on the survey – a question about whether firms used well-connected friends and family members to interact with government officials. We find no evidence that firms were underreporting socially undesirable behavior.

3.2 Sample

As is commonly the case in firm-level work, particularly in emerging markets, it was not possible to build a random sample of all the foreign-owned firms operating in the Philippines.⁹ Therefore, we constructed a convenience sample: the list was seeded based on firm lists from industry groups, government agencies, and chambers of commerce and then expanded via referrals from respondents. Firms were approached to participate in the study based on the following restrictions: 1) At least partially foreign-owned; 2) Located in Manila or Calabarzon;¹⁰ 3) Not a micro-enterprise.

With regard to external validity, we seek to infer that the patterns and relationships we observe within our sample of surveyed firms apply more broadly to the population of medium-large foreign firms operating in the Philippines. Our ability to do so is dependent on the degree to which the non-random sample of firms we survey is similar to that population overall. In the online appendix, we compare the firms in our sample to the 1224 largest multinational firms in the Philippines by size, sector, and home region. We discuss these comparisons in detail in Section B of the Online Appendix; we find a high degree of similarity between the firms in our sample and the total population of medium and large foreign firms operating in the Philippines.

The one dimension on which we expect our sample differs from the overall population is on a dimension we cannot observe for firms not in our sample, i.e. the degree of social connectedness. The disadvantage of our snowball sampling strategy is that firms with more firm-to-firm ties are more likely to end up in our sample. This is not a large concern in our study because our sample still contains enormous variation in level of connectedness and because we are exploring differences between specific types of social ties and the implications for political participation. However, it is a limitation on the external validity of our findings – our results may not inform us about firms with few or no peer connections as those firms are underrepresented in our sample.

3.3 Dependent variables

Our measures of political participation are two indicator variables that capture whether a firm attempted to influence policy at the local and/or national level.¹¹ Of the two levels of government, participation at the national level was more common, with 122 firms (56%) reporting an attempt, compared to 104 firms (48%) reporting an attempt to influence policy at the local level. Most firms attempted some form of policy influence: only 78 firms (35%) made no attempt at influencing policy at either level. A similar number (87 firms, or 39%) attempted to influence policy at both levels.

3.4 Independent variables

We focus on two types of social ties: ties to peer firms (peer ties) and ties to government officials (government ties). Because our sample is too small to be able to capture the entire network in which respondent firms are embedded, we instead gather information at the firm-level, asking managers about their firm’s relationships to different types of firms and different types of government officials.

3.4.1 Peer firms

We measure ties to peer firms by asking respondents the number of peer firms that they have a relationship with. The questionnaire was designed in order to differentiate between industry characteristics like industry concentration on the one hand and actual firm ties on the other, by inquiring about the peer firms with whom the firm in question has a relationship. Some firm-to-firm ties between peers are strictly organization-to-organization ties based on economic exchange between two firms in the same industry. However, many of these ties are social in nature, rooted in individual-to-individual relationships between owners and managers at one firm and owners and managers at another. These personal ties are formed through friendships, family relationships, or alumni or professional networks. Qualitative interviews with managers and firm owners identify their business school alumni networks as a particularly important source of connections.

To differentiate peer firms from clients, suppliers, or other partners, we asked about the number for firms in each category separately. Our variable uses the logged number of peer firms.

The module on firm social ties was divided into different categories by business function. Consequently, we introduced the module by inquiring about social ties of the firm and then asking for counts of these ties that are suppliers, clients, and peer firms. This particular set up was chosen because, during pretesting, respondents balked at having to consider social ties separately from the relationships that are central to business operations. When testing a module in which ties were divided as “social ties” vs. “business ties,” one of our respondents informed us that the task was not possible “because all of those business relationships are also social relationships.” While a different set-up for the question module and sharper delineation of business and social ties may be possible in other countries, we are operating in a context where there is little distinction between the professional and the personal.¹²

3.4.2 Political actors

Second, we focus on ties to political actors, in this case, either elected officials or bureaucrats. As in the peer ties, these relationships may form through a number of ways: friendships, family relationships, or professional networks. As our primary measure of government ties, we use as an indicator whether a member of the company leadership or the corporate board has held government or bureaucratic office–a relationship that denotes a substantial political relationship. In our sample, these ties are predominantly ties to the national legislature and the bureaucracy.

It is important to note that our measure of government ties captures a stronger political connection than asking if firms have social relationships with politicians (we capture and evaluate these weaker relationships in our alternative measure of frequency of interaction, described below). In the Philippines, almost all firms will have at least some social tie to politicians. In fact, even regular voters have a striking level of access to politicians–according to surveys in Cruz et al. (2017), for example, almost two-thirds of voters are either directly connected to the mayor or linked through one intermediary. Consequently, we focus on a measure of connectedness that captures a rarer type of connection – owners, managers, or board members who hold or have held office – in order to distinguish substantive ties to government institutions from the more common forms of social access to government officials.

3.4.3 Additional measures and checks

To assess the robustness of our main findings, we also use the frequency of interactions as an alternate measure of each tie type. The frequency of interactions with peer firms and government officials captures the actual behaviors that we would expect to be associated with firm networks. One way that firm connections operate is through providing opportunities for firms to engage with their peers and with government officials both formally and informally. Firm networks can provide contacts and opportunities to interact, and these interactions then facilitate attempts to work collectively with other firms to seek policy influence (in the case of peers), or to lobby directly for policy influence (in the case of government officials). Here, we use a survey question on the frequency of interactions with peer firms and with government officials. Interacts with Government takes the value of 1 if the firm reported interacting with government officials in the Philippines once a month or more, while Interacts with Firms uses the same measure for interactions with other firms.

One concern is that firms that are well-connected in one arena are similarly well-connected in the other, and that the only distinction that can be made empirically is between firms whose owners and managers are well-connected and those that are not. Fortunately, this is not the case. Our measures of government and peer ties are only weakly correlated with one another. We also show the government and peer ties among firms in our sample visually using a network graph in Fig. 1. Each of the dark gray circles (nodes) represents an actual firm in our sample (all 237 firms are displayed) and the light gray circles represent political actors: congress, local government, and the bureaucracy (labeled). For each firm, we show: (i) ties to Congress, bureaucracies, or local government depicted by connecting lines between the firm node and the government actor node; and (ii) ties to peer firms, represented by the size of the circle, where large circles are those firms that report a larger number of peer firms. A substantial number of firms, including many with a large number of peer ties, have no direct government ties. More broadly, Fig. 1 shows no distinct pattern in terms of number of peer firms and subsequent number of connections to the government or bureaucracy.

Fig. 1

Peer and government ties

Summary statistics are provided in Table 1. The exact wording of the questions on which our dependent and key independent variables are drawn is provided in the Online Appendix (Section D).

Table 1 Descriptive statistics

3.5 Model specification

To test our hypotheses, we use logistic regression to estimate the relationship between firms’ social ties and their attempts to influence government policy:

$$ \begin{array}{@{}rcl@{}} Pr(y_{f} = 1 | X,C) = P({\upbeta}_{1}X_{f} + {\upbeta}_{2}C_{f} + \sigma_{s} + \rho_{r} + \epsilon_{fsr} > 0) \end{array} $$

where y_f is a vector of indicator variables for firm f ’s attempts to influence policy: (i) AttemptLocal, indicating whether the firm has attempted to influence policy at the local level; and (ii) AttemptNational, indicating whether the firm has attempted to influence policy at the national level.

X_f refers to a vector of two variables indicating the type of ties that firm f has: (i) direct ties to bureaucrats or politicians; and (ii) the number of peer firms (logged).

Because firm-level factors also affect the political participation of firms (see, e.g., Weymouth (2012)), C_f represents size (in terms of revenues), age, profitability, and whether firm f is publicly traded. In addition, some models include controls for the number of employees, foreign ownership share, and whether firm f is located in a special economic zone.

Sector and regional dummies are represented by σ_s and ρ_r, respectively, while 𝜖_fsr is the usual idiosyncratic error term. Critically, the sector dummies capture all characteristics that are shared across firms in the same sector, including industry concentration.¹³

When comparing attempts to influence policy at the local vs. national level, we use seemingly unrelated regression to allow for the possibility that the errors are correlated across the different specifications (following Zellner (1962)). This correlation of errors would exist if firms’ decision to pursue local-level influence and their decision to pursue national-level influence are not independent, e.g. if focusing efforts to influence government at one level reduces the resources available to attempt to influence government at a different level.

4 Results

4.1 Social ties and political participation

We begin by evaluating the overall relationship between social ties and political participation, without distinguishing between local- and national-level politics. Table 2 presents the results of regressions that estimate the independent effects of peer ties and government ties on the probability that firms attempt to influence government policy at either the national or local level. We find that government ties are positively associated with these efforts (p<.05). In contrast, the estimated effect of peer ties is not significant in either specification, though the estimated effects are positive.

Table 2 Government ties, peer ties and attempts to influence policy (Pooled)

These findings are consistent with the existing literature, which has estimated positive effects of government ties on a variety of political outcomes for firms, while failing to find evidence that conditions conducive to collective action increase firms’ political efforts. However, these regressions don’t evaluate our theory directly. Doing so requires differentiating between political participation at the local and national level, which we undertake in Table 3.

Table 3 Government ties, peer ties, and political participation

4.2 Social ties and political participation at the local and national level

Table 3 reveals important differences in the value of different types of social tie for facilitation political participation at different levels of government. Consistent with H1, we find that peer ties are associated with attempts to influence policy at the national level. Consistent with H2a, government ties predict attempts to influence policy at the local level. These estimated effects are large. Holding all other variables constant¹⁴ and comparing a hypothetical firm in the 10th percentile for peer ties to one in the 90th percentile, we see a dramatic increase in the predicted probability of attempting to influence policy at the national level (from 14% to 81%). Similarly, we estimate that a hypothetical firm with no government ties has a 12% probability of attempting policy influence at the local level, while one with at least one tie has a 77% probability.

Recall that our regressions include a vector of dummy variables for sector, which absorb the effects of industry characteristics, including industry concentration and the extent to which an industry is heavily regulated. This is important because firms in heavily regulated industries may have greater incentives to seek political influence and firms in highly concentrated industries may have an easier time acting collectively. Thus, when we estimate that horizontal ties increase the likelihood that firms seek political influence at the national level, we are finding something above and beyond any relationship that might exist between industry characteristics and such participation.

The results in Table 3 are also consistent with H2b, the expectation that there is little or no relationship between peer ties and attempts to influence policy at the local level. We estimate a slightly negative, statistically insignificant, relationship between peer ties and attempts to influence policy at the local level. This suggests that reverse causation is not a likely explanation for our results – if politically active firms are pursued by other firms seeking peer ties, we would expect this dynamic to increase the number of peer ties for firms that are active at both the local and national level, not just firms active at the national level.

While we expect that the effect of government ties should be strongest at the local level, we also estimate a positive, though statistically insignificant, effect of government ties on attempts at the national level.¹⁵ Along with a similar and even stronger estimated effect in Table 4 below, this positive estimated effect likely indicates either that firms are sometimes willing to use their government ties in pursuit of political outcomes with broad benefits or that some firms are able to use government ties to seek particularistic benefits at the national level. This idea that firms can sometimes use direct ties to seek particularistic benefits at the national level is consistent with the existing literature on government ties, which highlights their value for things like securing government contracts or loans from state-owned banks (e.g. Malesky and Taussig 2009; Dinç 2005; Sojli and Tham 2017), both of which may be controlled at either the national or local level.

Table 4 Government interactions, peer interactions and political participation

4.3 Alternative measures of networks

One potential concern with the results in Table 3 is that these measures of government and peer ties are necessarily imperfect and may be noisy. One approach to address this issue is to use multiple measures that capture elements of the same concept, in this case, the frequency of a firm’s interactions with peer firms and government officials. If our results hold across both sets of alternative measures, it builds confidence that the underlying relationships are robust and not driven by any idiosyncracies of our measurement strategy.

Table 4 shows similar results using firms’ frequency of interaction with peer firms or government officials as an alternative measure of peer and government networks. The interactions with government actors are also significantly associated with attempts to influence policy at the national level. This is not inconsistent with our theoretical expectations. We expect that firms are more likely to be willing to bear the costs of leveraging government ties when they stand to reap all or most of the rewards of the policy change they seek, but there may still be circumstances when firms are willing to bear those costs in pursuit of policy changes that benefit a wide range of firms. It is also likely that, in some cases, firms seek particularistic benefits at the national level and use their government ties to do so. Thus, while we expect government ties to be most often used to pursue local-level policy change, the positive relationship between government ties and the likelihood of seeking policy influence at the national level is not surprising.

In Section A of the Online Appendix we demonstrate the robustness of our results to alternative specifications, including linear models and inclusion of both tie types in the same model.

4.4 Placebo test: Client ties

A core component of our theory is the claim that different types of social ties have different effects on firms’ political strategies. One additional way to test this empirically is to show that types of ties that we expect are irrelevant to political strategy are indeed poor predictors of attempts to influence policy. The ties between firms and their clients serve this purpose. While strong and numerous ties to clients are essential to business success, we do not expect them to facilitate collective action for political gain in ways similar to ties between peer firms.¹⁶ Thus, we expect client ties are poor predictors of political participation. The results in Table 5 confirm that client ties are indeed not significantly associated with attempts at influence.

Table 5 Client ties and participation at the local and national level

4.5 Falsification test: Firms in special economic zones

In addition to using other types of social ties in a placebo test, another approach is to demonstrate that peer firms and government ties do not affect outcomes that they would not be expected to affect. We can exploit differences among firms’ need for political engagement at either the local or national level to show that peer and government ties do not affect attempts to influence policy among firms with no incentive to do so. One convenient falsification test involves firms in special economic zones (SEZs) in the Philippines, because local governments are limited in their ability to tax or otherwise regulate firms in SEZs. Consequently, we expect that firms located in SEZs are much less likely to require engagement with local government and that they will primarily use their social ties to participate at the national level only.

If our data accurately capture patterns in the relationship between firms’ social ties and their political participation, we should see that for the subset of firms located in SEZs, social ties are only predictive of attempts to influence policy at the national level. Consistent with expectation, Table 6 shows that both government and peer ties are associated with participation at the national level for this subset of firms.

Table 6 SEZs: firm ties and participation at the local and national level

5 Conclusion

Peer ties are an important part of the social landscape of the firm, but the effect of these ties on firms’ political participation remains understudied. Peer ties bind groups of firms together, allowing social sanctions and social rewards and increasing firms’ ability and propensity to seek policy influence collectively on issue areas that affect large numbers of firms. We show that peer ties are associated with firms’ likelihood of pursuing policy change at the national level, where issues are relevant to sectors and larger groups of firms. Peer ties are not associated with pursuing policy change at the local level, where issues are generally narrower in scope, affecting only one or a handful of firms.

In contrast to peer ties, we expect that direct ties to government officials are primarily useful in the pursuit of narrow, particularistic benefits. We expect that firms prefer to bear the costs of calling on government ties in contexts where the scope of the policy change is narrow and they can consume most of the benefits themselves. Consistent with this expectations, we find evidence that ties to government officials are strong predictors of political participation at the local level. While we cannot rule out the possibility that government ties are also useful in seeking national-level influence, the effects we estimate here are smaller and not significant in most specifications.

Most of the policy debates related to firm lobbying focus on striking the right balance between allowing firms’ voice in government and over-representing their interests. However, in addition to the amount of influence firms have, it is equally important to understand how firms exercise their influence. The types of networking opportunities that governments foster can affect not only the extent of firms’ access, but also the fundamental way that firms engage with government. In particular, providing space for industry associations, chambers of commerce, and other firm associations may promote collective, rather than firm-specific modes of engagement with politicians and bureaucrats. In other words, if governments want to discourage requests for special treatment and exemptions at the firm level (and the accompanying bribes and facilitation fees), they can encourage collective action and industry-level engagement instead.

To return to the example of the mining industry in the Philippines, it may be more socially optimal for these firms to collective seek less restrictive environmental regulations than to individually seek non-enforcement of those regulations in specific instances. Even if collective lobbying seeks similar goals–for groups of firms rather than individual firms–collective participation is more consistent with conventional channels for lobbying and policy-making, making it easier for governments to integrate firm views in an appropriate manner.

Our work represents an important first step in understanding the distinct political effects of different types of social ties. While our results hold under alternative specifications and a variety of robustness checks, additional empirical work remains. Most importantly, full testing of our theory requires even more detailed measures of political participation that capture the precise issue areas over which firms seek influence. Additionally, it would be ideal to retest our theory in a wider range of countries and on domestic as well as foreign firms. At the same time, our work highlights the importance of a more nuanced approach to understanding the networks of firms. Different types of social ties provide firms with different capabilities, shaping the nature and extent of their political engagement.