Corporate Political Ties and Firm Performance in a Transition Economy: A Replication and Extension of Peng and Luo (2000)

Abstract

How do corporate political ties impact firm performance in a transition economy? This topic has attracted wide attention in the strategy field. Accordingly, our study replicates a highly influential study, ‘Managerial ties and firm performance in a transition economy: The nature of a micro-macro link’ (Peng & Luo, 2000). The original study found that managerial political ties greatly improve organizational performance, and that this ‘micro-macro’ link varies across ownership types, business sectors, firm sizes, and industry growth rates. This replication study offers a hierarchical view of political ties by extending it from the individual to the organizational level and explores the complex link between the two levels of corporate political ties and firm performance. The results of a staged quasi-replication exercise show some similarities with the original study in the mechanism of corporate political ties on firm performance but, more importantly, reveal some key differences in the effect size and contingent effects. Furthermore, an extended test shows that corporate political ties are multilevel, and different levels of political ties vary in their mechanisms and effects on firm performance. The findings reveal temporal and contextual sensitivities of political ties studies in transition economies.

摘要：

在转型经济体中，政治关联如何影响企业绩效？这一话题在战略管理领域引起了广泛关注。我们在本文中试图复制一项极具影响力的研究，即“转型经济中的管理层关系与企业绩效：一个‘微观——宏观’关系的本质”（Peng & Luo, 2000）。原研究发现，高管政治关联显著提高了组织绩效，而且这一“微观——宏观”关系因企业所有权性质、所属产业部门、企业规模及所处行业增长率不同而变化。本复制研究通过将政治关联从微观个体层面扩展到组织层面，引入政治关联的多层性观点，并全面检验两个层面的政治关联与企业绩效之间的关系。分阶段的准复制实证结果显示，虽然政治关联对企业绩效的影响机制与原始研究类似，但是这两项研究在关于政治关联效应大小和权变效应上存在关键差异。此外，进一步的拓展研究表明，企业政治关联是多层次的，不同层次的政治关联对企业绩效的作用机制和效应大小也存在显著差异。研究结果揭示了转型经济体中企业政治关联研究的时间和情境敏感性。

INTRODUCTION

Replication studies play an important role in advancing cumulative repeatable science (Bettis, Helfat, & Shaver, 2016; Tsang & Kwan, 1999). Top-tier publications, such as the Strategic Management Journal and Academy of Management Journal, have encouraged and called for replication studies in the management field (e.g., Bettis et al., 2016; Eden, 2002). Scholars have recognized that replication studies assist in building a cumulative body of knowledge and advance the rigor of research (Goldfarb & King, 2016) and have thus increasingly devoted themselves to conducting such studies in recent years (e.g., Bettis, 2012; Fang & Kim, 2018; Goldfarb, Zavyalova, & Pillai, 2018; Gupta, Mortal, & Guo, 2018; Mueller-Langerab, Fechercd, Harhoffb, & Wagner, 2019; Tsang, 2002; Tsang & Yamanoi, 2016).

Accordingly, we replicated an influential study published several years ago in the Academy of Management Journal, ‘Managerial ties and firm performance in a transition economy: The nature of a micro-macro link’ by Mike W. Peng and Yadong Luo (2000). This study conjectured that the social capital embedded in managerial (political) ties is important for firms operating in transition economies where firms face imperfect competition characterized by weak institutional support and distorted information (Peng & Luo, 2000: 486). Empirically, the study provided the first set of quantitative data demonstrating a positive and statistically significant association between managerial political ties and firm performance moderated by ownership type, business sector, firm size, and industry growth rate – in support of the central conjecture.^[1]

We chose Peng and Luo (2000) for the following three reasons. First, it was one of the first empirical studies to examine how political ties impact firm performance in a transition economy. Many papers have been published in prestigious journals since to boost related research (e.g., Arnoldi & Villadsen, 2015; Li & Zhang, 2007; Okhmatovskiy, 2010; Sheng, Zhou, & Li, 2011; Sun, Mellahi, & Thun, 2010; Sun, Peng, & Tan, 2017). Furthermore, the paper has become highly influential in the management field: as of August 2022, it had over 3,460 Google Scholar citations and over 1,700 citations in the Web of Science. In the five years up to 2022, the paper garnered over 1,300 Google Scholar citations and over 700 Web of Science citations and continues to gather further momentum (e.g., Guo, Shen, & Su, 2019; Hornstein & Zhao, 2018; Liedong & Rajwani, 2018; Zheng & Jin, 2018). Additionally, given the increasing prominence of transition and emerging economies, understanding how political ties impact firm performance is of topical importance. Overall, it is an exemplary study for replication.

Second, whether Peng and Luo (2000) – published more than 20 years ago – stands the test of time must be evaluated. Based on the perspective of political embeddedness, corporate political ties work by filling the institutional void (Ge, Carney, & Kellermanns, 2019). During this period, China has moved from the initial stage of a transition economy to the stage of comprehensive transformation and upgrading, and its economic and institutional background has undergone significant changes. For example, the dominant economic structure has shifted from manufacturing to services; the government–business relationship has become more diversified and open, and inter-organizational contacts have been gradually formalized. Based on the perspective of political embeddedness, it is doubtful whether political ties can still play a strong role and whether the way it works has changed in the current practices of Chinese enterprises, which encourages us to re-examine the conclusions of Peng and Luo (2000) by conducting replication studies.

Third, the design of Peng and Luo (2000) is not perfect. For example, Peng and Luo (2000) proposed and analyzed managerial political ties based on social network theory but overlooked the diversity and hierarchical nature of social relations. Following Peng and Luo (2000), scholars tend to view corporate political ties from a micro level (e.g., Burt & Opper, 2020; Guo, Xu, & Jacobs, 2014; Jia, Yi, & Zhang, 2019; Jiang, Guo, Wei, & Wang, 2018; Liedong & Rajwani, 2018; Wu, Johan, & Rui, 2016). Despite a few scholars’ attention to macro-level political ties (e.g., Inoue, Lazzarini, & Musacchio, 2013; Vaaler & Schrage, 2009), the two streams of research rarely engage in dialogue and integration (Sun, Mellahi, Wright, & Xu, 2015). Additionally, in their groundbreaking work, Peng and Luo (2000) used survey data to measure corporate political ties. However, this measure has inherent flaws, such as subjectivity and limited sample size. Moreover, subsequent studies have not unified the measure for political ties (or political connections) (e.g., Li & Zhang, 2007; Sheng et al., 2011; Sun et al., 2015; Zheng, Singh, & Mitchell, 2015). Therefore, it is necessary to replicate and extend Peng and Luo (2000) by adopting the improved method of measurement, so as to provide insights for contextualizing political connection research.

We chose a phased quasi-replication method, which refers to studies that assess robustness and/or generalizability using different measures or methods from the original study in stages (Bettis et al., 2016). First, we replicated the ‘micro-macro’ linkage model from Peng and Luo (2000) that tests the relationship between managerial political ties and firm performance. In this stage, we used data-mining and content analysis instead of the original study's questionnaire method to obtain objective data about managerial political ties and re-examined the micro–macro linkage. This allowed us to examine the robustness of the model across different measures and temporal contexts. Next, we used the same method to obtain data for organizational-level political ties and tested the macro–macro linkage between organizational political ties and firm performance. At this stage, to assess whether the original study's findings were robust across different levels, we broadened the hierarchy boundary of corporate political ties from the individual (managerial political ties) to the organizational level (organizational political ties). Finally, we extended Peng and Luo (2000) by conducting additional analyses to compare the effects of the two levels of corporate political ties on firm performance.

Our replication exercise makes two major contributions to the literature. First, we provide some non-findings, suggesting that the arguments of Peng and Luo (2000) are time- and context-sensitive. For example, our results show that the moderating role of the business sector is reversed, and that the explanatory power of corporate political ties on firm performance is much weaker than that in the original study. These non-findings confirm that the impact of corporate political ties on firm performance depends on the temporal and spatial context, thus providing insights for future research. Second, as an extension of Peng and Luo (2000), we provide a hierarchical view of corporate political ties and compare the impacts of different levels of political ties on firm performance. The results clearly show that managerial political ties exert a stronger influence on firm performance than organizational political ties when firms are under greater competitive pressures or facing a weaker institutional environment. However, with the expansion of firm size and the perfection of formal institutional environment, organizational political ties will play a greater role in explaining firm performance. These findings provide fresh insights into a more comprehensive understanding of corporate political ties, particularly in transition economies.

THEORETICAL BACKGROUND AND HYPOTHESES DEVELOPMENT

A Review of Peng and Luo (2000): Political Ties and Firm Performance in a Transition Economy

We start our replication study with a review of the theoretical arguments and hypotheses developed in the original study. The original study is motivated by the aphorism that managerial political capital is important to navigate environmental uncertainty. Peng and Luo (2000: 488) wrote that officials at various levels of the government had considerable power to approve projects and allocate resources in a transition economy like China. In such an environment, the state regulatory regime was the most influential, most complex, and least predictable environmental factor to impact firm performance (Peng & Luo, 2000: 488). The study thus postulated that arbitrary government intervention was a constant threat for numerous firms. To co-opt these sources of environmental uncertainty, business firms naturally maintained a disproportionately great deal of contact with government officials. Accordingly, executives with better ties with officials could navigate sources of environmental uncertainty better, thus effecting higher performance. The first hypothesis of the paper therefore states:

Hypothesis 1: Personal ties with government officials possessed by a firm's top managers are positively associated with firm performance.

Peng and Luo (2000) further pointed out that although political ties represent an important source of social capital, the diversity of Chinese organizations suggests that not all of them are interested in building political ties and that not all political ties are equally beneficial (Peng & Luo, 2000: 488). Firms at a competitive disadvantage tend to improve their performance by using political ties. First, compared with state-owned enterprises (SOEs), non-SOEs suffered from a lack of government support and their managers faced budgets and incentive structures that aligned closely with firm performance (Peng & Luo, 2000: 489). Thus, non-SOE managers were more strongly motivated than their SOE counterparts to improve firm performance by investing socially in political ties. Second, the impact of political ties on firm performance was stronger for service than manufacturing firms because around the year 2000, the service industry was just emerging, and the legal framework surrounding services was less developed than manufacturing in China's transition economy (Peng & Luo, 2000: 489). Moreover, manufacturing firms often needed to focus on developing technological and organizational capabilities, whereas service businesses were generally more relationship-intensive and depended more on external resources (Peng & Luo, 2000: 489). Third, smaller firms were more eager to build political capital to gain legitimacy and were more flexible and capable in establishing and improving business–government relations than bigger ones (Peng & Luo, 2000: 489). Finally, a slow-growth industry represented a saturated competing market. Consequently, firms might have been more strongly motivated to mobilize social ties to make a last-ditch effort to stay afloat and improve performance (Peng & Luo, 2000: 490). Thus, the original study developed a contingency framework by introducing the above firm- and industrial-level moderating factors and proposed the following hypotheses:

Hypothesis 2: The impact of political ties on firm performance is stronger for non-SOEs than for SOEs.

Hypothesis 3: The impact of political ties on firm performance is stronger for service than for manufacturing firms.

Hypothesis 4: The impact of political ties on firm performance is stronger for smaller than for larger firms.

Hypothesis 5: The impact of political ties on firm performance is stronger for firms operating in low-growth industries.

Theoretical Extension of Peng and Luo (2000): A Hierarchical View of Corporate Political Ties

Based on the social network theory, Peng and Luo (2000) pioneered the empirical study of political ties and firm performance from a micro–macro perspective. In their study, corporate political ties were measured as micro-level ties developed directly between firm executives and government officials. These ties imply managerial interpersonal connections with officials at various levels of the government (or the regulatory and supporting organizations) (Dong, Li, & Tse, 2013; Zheng et al., 2015). However, according to the social network theory, inter-organizational relationships including business–government ones are often multilevel in nature (Brass, Galaskiewicz, Greve, & Tsai, 2004; Sun et al., 2015; Zaheer, Gozubuyuk, & Milanov, 2010). Defined as ‘boundary-spanning personal and institutional linkages between firms and the constituent parts of public authorities’ (Sun, Mellahi, & Wright, 2012: 68), corporate political ties are generally a widespread organizational arrangement and a crucial component of corporate non-market strategy (Durand, Grant, Madsen, Dorobantu, Kaul, & Zelner, 2017; Faccio, 2006). Thus, corporate political ties should embrace both personal-level (e.g., Liedong & Rajwani, 2018; Peng & Luo, 2000) and organizational-level (e.g., Inoue et al., 2013; Vaaler & Schrage, 2009) connections.

Political connections at the organizational level exist in different forms (Sun et al., 2010). State ownership ties are natural, formal, and stable inter-organizational relationships, which often come at the expense of corporate autonomy (Bortolotti & Faccio, 2009; Okhmatovskiy, 2010; Sun et al., 2015). Several scholars regard state ownership as an institutional arrangement and consider it an independent element. For example, some studies suggest that state ownership is an important contingency factor influencing the relationship between political ties and firm performance (e.g., Liu, Lin, & Wu, 2018; Peng & Luo, 2000). However, per social network theory, organizations, like individuals, may also build political ties through active behaviors such as political strategies, especially among firms without a natural institutional connection (Marquis & Qian, 2014; Oliver & Holzinger, 2008). For example, firms might actively participate in and facilitate government orders or projects (McDonnell & Werner, 2016; Oliver & Holzinger, 2008), apply for political patronage (Faccio, Masulis, & McConnell, 2006), or invite government officials to inspect them, and signal this to the outside world to enhance their legitimacy (Ridge, Hill, & Ingram, 2018; Sun et al., 2010). Unlike in state ownership, focal firms usually acquire these political ties through political initiative and strategic behaviors. These strategic political ties play an important role in firm performance (Oliver & Holzinger, 2008). To conduct a dialogue with Peng and Luo (2000), we introduce organizational-level political ties from the social network and a strategic initiative perspective, adopting a hierarchical view to review the relationship between corporate political ties and firm performance.

According to theoretical arguments from the original study, not all corporate political ties are equally beneficial to all firms (Peng & Luo, 2000). From the perspective of social network theory, the multiplicity of business–government relations suggests variations in corporate political ties developed by organizations and individuals (Sun et al., 2015). For example, the costs of maintaining two types of corporate political ties differ. Managerial political ties, often based on social exchanges via interpersonal relations, require more social emotional investments such as trust rather than financial costs (Uzzi, 1997). In contrast, organizational political ties, which usually result from corporate political strategies (Durand et al., 2017; Oliver & Holzinger, 2008), require more financial costs. Additionally, their working mechanisms on firm performance differ. As an informal embedded relationship, the mechanism of managerial political ties seeks to help the focal firm obtain more scarce resources and information at a low financial cost to promote performance, especially when the formal institutions are imperfect (Peng & Luo, 2000). Meanwhile, organizational political ties improve firm performance mainly by aligning the strategic goals and incentives between the focal firm and the government and sending the outside world a signal of legitimacy (Ridge et al., 2018; Sun et al., 2010). Based on social network theory, in the Chinese transition economy's environment, where formal institutions are rather weak, embedded interpersonal connections provide greater rent-seeking opportunities by nurturing greater emotional trust, commitment, and reciprocity (Fang, Chi, Chen, & Baron, 2015; Ge et al., 2019; Guo, Tang, & Wei, 2020; Peng, 2003; Peng & Heath, 1996). In view of this, we argue that managerial political ties are more important than organizational political ties in the context of the transition economy, as the immature institutional environment generally leaves firms with less institutional support and/or a stronger incentive to mobilize social ties to make a last-ditch effort to stay afloat (Peng & Luo, 2000).

In line with Peng and Luo (2000), we suggest that the difference in the benefits from these two types of ties is influenced by the diversity of organizations and the special institutional environment in China as well. Furthermore, non-SOEs, manufacturing firms, smaller firms, and firms in lower-growth industries face greater institutional voids and rent-seeking motivation, such that the managerial political ties provide greater rent-seeking opportunities. Recent studies have highlighted that the marketization level, which is a unique institutional factor in the context of the Chinese transition economy, is an important factor affecting the mechanism of corporate political ties (e.g., Wang & Qian, 2011; Zhou, 2014). China has transitioned from a planned to a market economy, and the level of marketization varies significantly among regions. In the case of a low-marketization level, market maturity is low, and the legal system is ineffective, and management political ties play a stronger role in filling the institutional voids than organizational ties. With improved marketization levels, formal institutions tend to be perfected, and the rent-seeking advantage of managerial political ties gradually fades (Sun et al., 2015). Thus, we hypothesize the following:

Hypothesis 6: The impact of managerial political ties on firm performance is stronger than that of organizational political ties, especially for non-SOEs, manufacturing firms, smaller firms, firms in lower-growth industries, and firms in low-marketization regions.

METHODS

Replication studies vary according to the data sources, measures of construct, and analytical methods (Tsang & Kwan, 1999). Bettis et al. (2016) divided replication research into two broad categories: (1) narrow replication using the same data (or different data, but drawn from the same population) and research design as in the original study and (2) quasi-replication that assesses whether the results of an earlier study hold when using different research designs. In this study, we adopt the quasi-replication method for the following reasons. First, quasi-replication plays an important role in the field of strategic management because it informs us about how well results hold up in multiple settings using different measures and methods (Bettis et al., 2016). Second, Peng and Luo (2000) used questionnaire survey data to test their model. Although we believe that this has contributed to empirical studies of corporate political ties and boosted research in the field, as they acknowledged, this practice has certain flaws, such as subjectivity, that prevent us from performing strict narrow replication.

To understand whether, how, and why the results of the original study hold up, altering the measure, method, and model in different replication stages in order is important (Bettis et al., 2016). Our replication thus takes a three-stage sequential approach: in the first stage, we focus on the micro–macro link between managerial political ties and firm performance as Peng and Luo (2000) did. However, we move beyond the original work's constraints by generating a sample that covers a longer period and by developing objective measures for political ties with the help of data-mining technology. In the second stage, we further test the macro–macro link between organizational political ties and firm performance, which should enhance confidence in our findings from the previous stage. Finally, to extend the work of Peng and Luo (2000), we conduct further analysis to compare the impact of the two different levels of political ties on firm performance and explore their interplaying effect.

Sample and Data

We selected samples with reference to Peng and Luo (2000), with a more comprehensive consideration. Specifically, our sample came from listed companies in five megacities in China – Beijing, Shanghai, Tianjin, Chongqing, and Shenzhen – with data gathered from 2010 to 2015. We chose these five cities for the following reasons. First, they serve as important examples of China's reform and opening up in the current transition period, just as Peng and Luo (2000) selected five coastal provinces and one pilot province with reform and opening up. They effectively reflect the characteristics of economic transformation in respective periods. Second, their political structures are similar and comparable, as they are essentially provincial-level entities in China's political system. Specifically, Beijing, Shanghai, Tianjin, and Chongqing are special municipalities, and Shenzhen is China's first special economic zone. Finally, the sample is regionally balanced because the five cities are in different regions of China: Beijing and Tianjin are in Northern China, Shenzhen in Southern China, Shanghai in Eastern China, and Chongqing in Western China. They represent the institutional environments of different regions of China, reflecting the diversity of the sample sources. For example, Chongqing's marketization level is much lower than Shanghai's.

Our dataset was based on multiple sources. The data related to corporate political ties was obtained from media coverage (i.e., government websites, official websites, and annual reports of firms). To code the data, we first built a list of political actors and governmental agencies in the five specific cities during 2005–2015, referring to resolutions drafted by the People's Congresses of various provinces, which is public information on government websites.^[2] According to the list above, we used data-mining to obtain all news texts containing the information of both the listed company's name (or executive's name) and the political actor's name (or governmental agencies) from firm reports, which is public information on government and firm websites (Chakrabarti, 2003); thereafter, we captured and analyzed the information according to the principles in Table 1. After removing duplicate information and data, we obtained 92,135 pieces of information that indicated the existence of corporate political ties. Next, we identified and calculated the yearly number of corporate political ties between political entities and 438 listed companies in the five cities above for a total of 2,448 observations.

Table 1. Corporate political ties

Other firm-level information was obtained from the WIND and CSMAR databases of China. Industry-level information was aggregated from firm-level information.

Measures

Independent variables

First, we analyzed news content and extracted managerial political ties according to two criteria (shown in Table 1): The political actors had worked in focal firms (Sun et al., 2017), or the CEO had worked together with political actors in the same department. We calculated the degree of a firm's managerial political ties according to the measure of actor degree centrality per Freeman (1979). Thus, the degree of a focal firm's managerial political ties is defined as follows:

(1)$$\qquad\qquad {\rm Managerial}\,{\rm political}\,{\rm tie}{\rm s}_{\, i} = \displaystyle{{\mathop \sum \nolimits_{\,j = 1}^g x_{ij}\left( {i\ne j} \right)} \over {g-1}}$$

Here, j represents the central and local government actors listed; x_ij = 1 if and only if the focal firm i and political actor j are connected by a line, and 0 otherwise; g is the size of the political network group including the focal firm and political actors (g − 1 represents the maximum number of possible connections the focal firm can develop with political actor nodes). We calculated this measure on a yearly basis for every listed company using their number of personal connections with local political leaders (Oehme & Bort, 2015).

Second, based on the existing literature (Faccio et al., 2006; Marquis & Qian, 2014; McDonnell & Werner, 2016; Oliver & Holzinger, 2008), we analyzed news content and extracted organizational political ties according to the following three criteria, as shown in Table 1: (1) General leaders and deputy leaders of the government inspected the firm or made speeches related to the company. (2) The government signed agreements such as a government procurement agreement with the company or issued certificates to the company. (3) The company was supported or subsidized by the government. Furthermore, we calculated the degree of a firm's organizational political ties according to the measure of actor degree centrality per Freeman (1979). Thus, the degree of a focal firm's organizational political ties is defined as follows:

(2)$$\qquad\qquad {\rm \;Organizational}\,{\rm political}\,{\rm tie}{\rm s}_{\, i} = \displaystyle{{\mathop \sum \nolimits_{\,j = 1}^g x_{ij}( {i\ne j} ) } \over {g-1}}$$

Here, j is the governmental agencies listed; x_ij = 1 if and only if the focal firm i and governmental agency j are connected, and 0 otherwise; g is the size of the political network group including the focal firm and governmental agencies (g − 1 represents the maximum number of possible connections the focal firm can develop with governmental agencies).

Dependent variable

Following Peng and Luo (2000), we measured firm performance by return on assets (ROA). We used one-year lagged ROA as the dependent variable, which thus ranged from 2011 to 2016.

Moderating variables

Per Peng and Luo (2000), we introduced four moderating variables. Ownership was measured by a dummy variable, coded as 1 for SOE and 0 for non-SOEs. Business sector was measured by dummy variables 1 for manufacturing and 0 for service. Firm size was defined by the number of employees; we used the logarithm of the number of employees to make the variable more stationary. Industry growth was measured by the compound growth rate of pretax profits. Additionally, we introduced the marketization index for China's provinces (2010–2016) to capture the changes in local marketization levels (Wang & Qian, 2011).

Control variables

Based on existing research, we controlled seven firm-level variables and three industry-level variables. Specifically, firm age was determined by the founding year of the firm (Peng & Luo, 2000). The prior performance was measured by sales growth. Board size was the number of a firm's directors. Board independence was measured by the number of independent directors divided by the total number of directors. CEO duality was a dummy variable coded as 1 if the CEO also served as board chair, and 0 otherwise (Boyd, 1995). CEO tenure was measured by counting the months a chief executive had been in office (Henderson, Miller, & Hambrick, 2006). Ownership concentration was measured by the shareholding ratio of the largest shareholder (Li, Guo, Yi, & Liu, 2010b). Industry munificence was measured by the growth in industry sales over a five-year period, that is, the regression slope coefficient based on the regression of time against the value of shipments, divided by the mean value. The estimate for any given year is based on the five preceding years, that is, the munificence estimate for 2010 is based on 2006–2010 data (Dess & Beard, 1984). Industry dynamism was measured by the standard error of the regression slope coefficient divided by the mean value using the same regression model as for industry munificence (Dess & Beard, 1984). Industry competition was measured by the industrial Herfindahl index (Li, Poppo, & Zhou, 2010a).

RESULTS

Replication Stage One: Managerial Political Ties and Firm Performance

Hypotheses tests

In line with Peng and Luo (2000), we re-examined the micro–macro link between managerial political ties and firm performance. Given that the sample firms were from five different megacities, this could lead to clustered errors. We first conducted the zero-model test to compute the intraclass correlation coefficient (ICC) (Hox, Moerbeek, & Schoot, 2017) as follows:

$$\qquad\qquad\quad \rho _I = \displaystyle{{\tau ^2} \over {\tau ^2 + \sigma ^2}} = \displaystyle{{0.00015} \over {0.00015 + 0.26597}} = 0.00056$$

where τ ² is the population variance between clusters and σ ² is the population variance between clusters.

The results show no substantial ICC (ρ < 0.001), which means that the samples in different cities had almost no clustered errors. In such cases, where the observations are nearly independent, traditional multiple regression analysis will provide accurate estimates of the parameters and standard errors (Hox et al., 2017). We thus normalized the independent variable and tested our hypotheses via multiple regression analysis as Peng and Luo (2000).

Table 2 presents the means, standard deviations, and correlations among the variables examined in this study. The correlation of managerial political ties and firm performance in our study (ρ = 0.217, p < 0.001) is similar to that of the original study (ρ = 0.32, p < 0.001).

Table 2. Descriptive statistics and correlation matrix

Notes: Standard errors are reported in the parentheses.

* **, and *** denote significance at the 5%, 1%, and 0.1% levels, respectively.

Table 3 and Table 4 present the regression results for testing hypotheses of Peng and Luo (2000). Table 5 juxtaposes the main results of this replication study with the results reported by Peng and Luo (2000). For testing Hypothesis 1, the results show that managerial political ties are positively related to firm performance (β = 0.134, p < 0.001), which is consistent with the original study's results (β = 0.24, p < 0.05).

Table 3. Results of multiple regression analysis: Test for main effects

Notes: Coefficients of city and year dummies are not reported. Standard errors are reported in parentheses.

+ *, **, and *** denote significance at the 10%, 5%, 1%, and 0.1% levels, respectively.

Table 4. Results of multiple regression analysis: Test for moderating effects

Notes: Coefficients of city and year dummies are not reported. Standard errors are reported in parentheses.

+ *, **, and *** denote significance at the 10%, 5%, 1%, and 0.1% levels, respectively. Z-statistics are reported in square brackets.

Table 5. Comparisons of empirical results reported by the original study and the replication study

Notes: Coefficients of city and year dummies are not reported. Standard errors are reported in parentheses.

+ *, **, and *** denote significance at the 10%, 5%, 1%, and 0.1% levels, respectively.

Table 4 reports a series of subgroup analyses testing the contingency hypotheses (H2–H5). Specifically, for testing Hypothesis 2, the original result shows that managerial political ties are significantly and positively associated with firm performance for non-SOEs (β = 0.20, p < 0.05) yet not associated with performance for SOEs (β = 0.08, n.s.). Our subgroup analysis shows that the impact of managerial political ties on performance is stronger for non-SOEs (β = 0.213, p < 0.001) than for SOEs (β = 0.009, p < 0.001). To determine whether this difference is statistically significant, we calculated the Z-statistic for managerial political ties. The Z-statistic (9.671) rejects the null hypothesis of the coefficients being equal at the 0.1% level (Clogg, Petkova, & Haritou, 1995), which suggests that the effect difference is significant.

In testing Hypothesis 3, the original study's result shows that the impact of managerial political ties on firm performance is stronger for service firms (β = 0.22, p < 0.01) than for manufacturing firms (β = 0.08, n.s.). However, contrary to the original result, our subgroup analysis shows that the impact is stronger for manufacturing firms (β = 0.171, p < 0.001) than for service firms (β = 0.070, p < 0.001), and the Z-statistic (4.249) rejects the null hypothesis of the coefficients being equal at the 0.1% level.

Consistent with Peng and Luo (2000), our subgroup analysis for Hypothesis 4 shows that managerial political ties are significantly and positively associated with performance for smaller firms (β = 0.24, p < 0.01 in the original study; β = 0.203, p < 0.001 in this replication study) yet not associated with performance for larger firms (β = 0.08, n.s., in the original study; β = 0.003, p < 0. 1, in this study), and the Z-statistic (8.663) rejects the null hypothesis of the coefficients being equal at the 0.1% level.

Results from the original study suggest that managerial political ties are important for both firms in low- and high-growth industries (β = 0.20, p < 0.05 for both industries) failing to support Hypothesis 5. Yet, our replication result shows that the impact of managerial political ties on performance is stronger for firms in low-growth industries (β = 0.169, p < 0.001) than for firms in high-growth industries (β = 0.077, p < 0.001), and the Z-statistic (3.740) rejects the null hypothesis of the coefficients being equal at the 0.1% level.

Furthermore, there are significant differences in models that explain variances between the replication study and the original study. As shown in Table 5, adjusted R ² reported in the replication study is much smaller than that in the original study. The huge unexplained variance in the replication study means that the positive effect size of managerial political ties on firm performance is very small. Therefore, we have not been able to replicate Peng and Luo's (2000) conclusion exactly.^[3]

Robustness checks

We conducted several robustness checks to ensure the credibility of our results. First, we re-examined all five hypotheses with fixed-effects models (Tables 6 and 7). The results remain qualitatively unchanged. Second, the models may have endogeneity problems. For example, not all firms can build political ties, and various factors may influence the likelihood of building political ties and firm performance simultaneously, thus causing a self-selection bias. To address the endogeneity problem, we first controlled for the performance of the previous period in the model and obtained similar results. Furthermore, we conducted a coarsened exact matching (CEM) analysis to further assess the robustness of our findings (Blackwell, Iacus, King, & Porro, 2009; Iacus, King, & Porro, 2012). We began by categorizing managerial political ties as either relatively low or high level using the mean value of managerial political ties as the cutoff point. Thereafter, we identified matched observations that were similar based on firm size, ownership, and marketization index but different in terms of managerial political ties. In the control group, 1,519 of the 1,574 observations matched; the number was 860/874 in the treatment group. Multivariate L1 distance reduced from 0.37 to 0.30, and the imbalance on the other quantiles also improved, indicating that our match was successful. We then conducted two feasible ordinary least squares (OLS) regressions on the combined samples of observations with low and high ties using the CEM procedure. The results from both regressions were consistent with findings from other models.

Table 6. Results of fixed-effect models: Test for main effects

Notes: Coefficients of city and year dummies are not reported. Standard errors are reported in parentheses.

+ *, **, and *** denote significance at the 10%, 5%, 1%, and 0.1% levels, respectively.

Table 7. Results of fixed-effect models: Test for moderating effects

Notes: Coefficients of city and year dummies are not reported. Standard errors are reported in parentheses.

+ *, **, and *** denote significance at the 10%, 5%, 1%, and 0.1% levels, respectively. Z-statistics are reported in square brackets.

Replication Stage Two: Organizational Political Ties and Firm Performance

Hypotheses tests

In this stage, we considered a macro alternative measure of the independent variable to examine whether the original findings are consistent across different levels. Specifically, we tested the macro–macro link between organizational political ties and firm performance to re-examine the original arguments.

For testing Hypothesis 1, Model 3 of Table 3 shows that organizational political ties are positively related to firm performance (β = 0.088, p < 0.001). Table 4 reports a series of subgroup analyses testing the contingency hypotheses (H2–H5). The results show that the impact of organizational political ties on performance is stronger for non-SOEs (β = 0.113, p < 0.001) and smaller firms (β = 0.138, p < 0.001) than for SOEs (β = 0.010, p < 0.001) and larger firms (β = 0.006, p < 0.001), and the Z-statistics (5.124 and 4.711, respectively) reject the null hypothesis of the coefficients being equal at the 0.1% level, thus supporting Hypotheses 2 and 4. Contrary to the original result for testing Hypothesis 3, we find that the impact of organizational political ties on firm performance is stronger for manufacturing firms (β = 0.093, p < 0.001) than for service firms (β = 0.042, p < 0.001), and the Z-statistic (2.232) rejects the null hypothesis of the coefficients being equal at the 5% level, thus confirming our findings in Stage One. Testing Hypothesis 5, our replication results in Table 4 show that the impact of organizational political ties on performance is stronger for firms in low-growth industries (β = 0.088, p < 0.001) than for firms in high-growth industries (β = 0.046, p < 0.001), and the Z-statistic (1.780) rejects the null hypothesis of the coefficients being equal at the 10% level, thereby supporting Hypothesis 5.

However, like the results of Replication Stage One, there are huge unexplained variances in this replication stage, which means that the positive effect size of organizational political ties on firm performance is small.

Robustness checks

We conducted the same robustness checks as in Stage One and obtained similar results. First, we re-examined all five hypotheses with fixed-effects models (Tables 6 and 7). The results remain qualitatively unchanged. To address the self-selection bias discussed earlier and the endogeneity problems that might arise when government officers tend to visit high-performance firms, we first controlled for the performance of the previous period in the model and obtained similar results. Second, we conducted the CEM procedure as did in Replication Stage One, and the result was robust. Specifically, in the control group, 1,417 of the 1,495 observations matched; the number was 945/953 in the treatment group. Multivariate L1 distance reduced from 0.372 to 0.287, and the imbalance on the other quantiles also improved. We then conducted OLS regression on the combined samples of observations with low and high ties using the CEM procedure and obtained similar results.

Extension Study: Corporate Political Ties’ Heterogeneity and Firm Performance

To test Hypothesis 6, we conducted further analyses to compare the impact of the two types of political ties on firm performance for the full sample and subgroups. First, we compared the coefficient difference between managerial political ties and organizational political ties in Model 4 of Table 3 (with the full sample). As shown in Model 4, the impact of managerial political ties on firm performance (β = 0.127, p < 0.001) is stronger than that of organizational political ties (β = 0.075, p < 0.001). We then conducted the Wald test to verify the validity of the coefficient difference (see Table 8). The results show that managerial political ties have a stronger impact on firm performance (F = 8.06, p < 0.01). Second, we verified the validity of the coefficient differences between managerial and organizational political ties by subgroup analyses. As shown in Table 4, all coefficients of managerial political ties are larger than that of organizational political ties in the subgroup analyses of non-SOEs, manufacturing firms, smaller firms, and firms in lower-growth industries, and in areas with low-marketization levels. The results of the Wald tests confirm that the impact of managerial political ties on firm performance is significantly stronger than that of organizational political ties for the subgroups (see Table 8).

Table 8. Comparisons of coefficient differences

Notes: Based on the regression results in Tables 3 and 4. Standard errors are reported in parentheses.

+ *, **, and *** denote significance at the 10%, 5%, 1%, and 0.1% levels, respectively. F-values in the fixed-effect models are reported in square brackets.

However, in the other five subgroups, managerial political ties did not all show an advantage. Especially in the subgroups with larger size and a high-marketization level, the coefficients of organizational political ties are larger than that of managerial political ties, and the coefficient differences are supported by the Wald test (F = 2.82, p < 0.1; F = 7.01, p < 0.01, respectively). This reveals that the impact of organizational political ties on firm performance is stronger than that of managerial political ties when the firm size is large or the level of marketization in the region where the firm is located is high.

Overall, Hypothesis 6 is supported by our data. Additionally, we conducted robustness checks with the fixed-effects models to ensure the credibility of our results. The results remain qualitatively the same (see Table 8).

DISCUSSION

Comparing the Replicating Findings with Peng and Luo (2000)

This study has revisited the relationship between corporate political ties and firm performance examined in the seminal work of Peng and Luo (2000). The results of a staged quasi-replication exercise show some similarities with the original study in the mechanisms of corporate political ties on firm performance but, more importantly, reveal some key differences.

Changes in economic structure and contingent effects of political ties

Peng and Luo (2000) suggested that the impact of political ties on firm performance was stronger for service than for manufacturing firms, because the service industry was just emerging around the year of 2000. In Peng and Luo (2000), as shown in Table 5, there were only 28 service firms in the total sample of 127. Moreover, the legal framework surrounding the service industry was less developed in China's transition economy (Peng & Luo, 2000: 489). However, the replication study suggests the exact opposite. A key reason is that China's economic and industrial structure have undergone a slew of reversals in the past 20 years. After 20 years of development, China's economy now ranks second in the world, and its GDP has expanded tenfold. The service sector in China has been growing rapidly and shaping a gradually perfect industrial environment, while the manufacturing sector is facing growing challenges under the trend of deindustrialization. Consequently, manufacturing firms may have to invest more heavily in political ties to co-opt sources of environmental uncertainty in a bid to stay afloat and improve performance.

Development of institutional environment and effect size of political ties

Regarding the explanatory strength of the empirical models, our replication exercise exhibits high unexplained variance, which is an important non-finding compared with the original study. In addition to sample size limitations of Peng and Luo (2000) and differences in measurement methods between the replication and original study, a more critical reason might be the development of China's market and institutional environment.

First, Peng and Luo's (2000) conclusion presupposes imperfected competitions characterized by weak institutional support and distorted information. After 20 years of development and transformation, China's institutional context has fundamentally changed. The level of marketization in China has gradually improved in the past 20 years, which is mainly reflected in significant improvement in the relationship between governments and market, the development of non-state economy, product market, factor market, the development of organizations in the market, and the legal environment (Fan, Ma, & Wang, 2019; Wang, Fan, & Yu, 2021). Based on the political embeddedness perspective, the role of political ties depends on the perfection of formal institutions (Brockman, Rui, & Zou, 2013; Liu et al., 2018; Okhmatovskiy, 2010; Sun et al., 2010). Filling the institutional void is an important way for corporate political ties to play a rent-seeking role. With improved formal institutions, the rent-seeking ability of corporate political ties is gradually weakened (Zhou, 2014).

Second, the benefits of corporate political ties are vulnerable to political events (Sun et al., 2015). The Chinese political environment has changed dramatically in recent years, especially since Xi Jinping came to power. Anti-corruption efforts have reduced the rent-seeking ability of corporate political ties (Liu et al., 2018; Yu, Zhang, Tan, & Liang, 2022). Additionally, with the complexity of organizational forms and environments, the factors influencing firm performance become more diversified, and the explained variance of empirical models in recent research shows a weak trend (e.g., Brockman et al., 2013; Liu et al., 2018). Therefore, these non-findings suggest that the results of Peng and Luo (2000) are time- and context-sensitive.

The hierarchy of corporate political ties and heterogeneity between their effects

Peng and Luo (2000) focused on the value of personal (political) ties on firm performance, which is based on the imperfected competition and the unique guanxi culture of China. For a long time, personal relationship (guanxi) was one of the major dynamics in the Chinese society and a pervasive part of the Chinese business world. It binds literally millions of Chinese firms into a social and business web (Chen, Chen, & Huang, 2013; Luo, 1997). It was widely recognized that personal political relationship is a key business determinant of firm performance (Li & Zhang, 2007; Luo, 1997; Okhmatovskiy, 2010; Peng & Luo, 2000; Sheng et al., 2011). However, with the rapid development of China's market economy and continuous improvement of formal institutions, organizational relationships have been changing from relying on personal relationships to formal and diversified directions (Sun et al., 2012). In the extension exercise, combining social network theory, we introduced a hierarchical view of corporate political ties by extending micro- to macro-level political ties. Our findings show that the impact of managerial ties is stronger than that of organizational political ties when firms are under greater competitive pressure or weaker institutional environments. With the expansion of firm size and perfection of formal institutions, organizational political ties will play a greater role in improving performance.

Contributions

Our replication and extension work contributes to the field in two primary ways. First, the replication study provides additional statistical evidence and non-findings for the value of corporate political ties by performing a quasi-replication of Peng and Luo (2000) by altering the measures, methods, and design models in order in different stages – as called for in replication studies (Bettis et al., 2016). Peng and Luo (2000) suggested that corporate political ties strongly help to improve firm performance, and this impact is stronger for non-SOEs, smaller firms, and firms from low-growth industries. In the replication exercise, we provided evidence for the time sensitivity of the original study by casting some doubt on the moderating effect of the business sector and the huge unexplained variance of the empirical models. The replication exercise shows that corporate political ties’ impact on firm performance is stronger for manufacturing firms than for service firms, and more importantly, all the models in the replication study show large unexplained variances, in contrast to Peng and Luo (2000).

These results indicate that the effectiveness of corporate political ties decreases as the economic and institutional environment improves. In other words, our non-findings are another way of supporting the key argument of Peng and Luo (2000) that the social capital embedded in political ties may be more important in imperfect competition characterized by weak institutional support and distorted information. Therefore, our work provides compelling evidence that temporal consideration and the institutional change behind it may be important boundary conditions for understanding the firms’ political activities and their outcomes. This finding offers an opportunity for a more comprehensive understanding of the dynamic impact of corporate political ties on firm performance.

Second, we extend Peng and Luo (2000) by introducing a hierarchical view of corporate political ties, comparing micro- and macro-level corporate political ties and their effects on firm performance. Recognizing that network relations at different levels differ in nature (Brass et al., 2004; Zaheer et al., 2010), social network scholars have called for further studies to differentiate hierarchical network relations (Sun et al., 2012, 2015). Regarding the linkage between corporate political ties and firm performance, however, there is a long tradition among scholars to observe corporate political ties from a micro level (e.g., Liedong & Rajwani, 2018; Peng & Luo, 2000). Few scholars pay attention to the macro level, especially from the perspective of organizational initiative. Furthermore, the two levels of political ties in existing research are largely independent, lacking dialogue and integration (Sun et al., 2015), because the two ties differ in characteristics, maintenance cost, and their influencing mechanism on firm performance.

In this study, we compare the impact of two types of corporate political ties on performance. Our results show that the impact of managerial political ties is stronger than that of organizational ones, especially for firms at a competitive disadvantage or under weak institutions. In a transition economy, where the formal institution framework is weak, interpersonal social ties show a stronger rent-seeking ability by providing quality and timely information, resources, and by facilitating transactions at lower maintenance costs. As the formal institution improves, this advantage might gradually disappear, formal relations at the organizational level will gain greater legitimacy. These findings provide fresh insights into the systematic understanding of corporate political ties.

Limitations and Future Research Directions

This study has some limitations. Like Peng and Luo (2000), our replication work is limited to local areas in China. There exist diverse subcultures in China, and the institutional environments vary a lot among regions. Future research based on more extensive regional coverage might reveal a more insightful and comprehensive picture of the influence of political ties on firm performance in emerging markets.

Our replication and extension work offers several promising avenues for future research. First, scholars must pay attention to corporate political ties’ heterogeneity by testing the roles of different levels of corporate political ties on firm value. Although our study provides a snapshot of how different political ties impact firm value, the underlying mechanism should be further analyzed. On the one hand, future studies can contribute to the political strategy literature by achieving a deeper understanding of the process of how different political tie configurations impact firm performance. On the other hand, as our results suggested, different types of corporate political ties may be intertwined, overlapping, or interactional (Sun et al., 2015). Future work on corporate political ties may look for ways to offer fresh insights into the complexity around these relationships, such as distinguishing between their separate effects and understanding the interplaying effect.

Additionally, our study shows that with the improvement of formal institutions and the complexity of organizational forms and environments, the explanatory power of corporate political ties over firm performance decreases. Future work can focus more on the contingent value of these different types of political ties by exploring their boundary conditions. For example, firms at different stages of development or in different industrial, market, and institutional environments may have different appeals for different political tie configurations. A high-profile political shock in the transition economy may also have far-reaching implications for the value of different types of corporate political ties (for recent examples, see Liu et al., 2018; Sun et al., 2015). Investigating these questions requires further contingency studies.