Introduction
Research on state-owned enterprises (SOEs) has established that state ownership gives these firms distinct governance mechanisms and goals that orient them to carry out state missions and strategies (Huang & Yu, Reference Huang and Yu2006). However, the state is not a monolithic entity, as revealed in a growing number of studies. For instance, the central and local governments can have divergent interests and goals (Luo, Wang, & Zhang, Reference Luo, Wang and Zhang2017). Examining such differences is critical for understanding the implementation of national strategies concerning the economy and corporations. To better understand the state's role in corporate change, prior research has started to analyze administrative capacity and state fragmentation (e.g., Chung, Luo, & Zheng, Reference Chung, Luo and Zheng2024; Guillén & Capron, Reference Guillén and Capron2016). However, research lacks on whether and why SOEs owned by different government entities respond differently to national policies.
To fill this research gap, this study draws on the attention-based view (ABV) to examine how central SOEs and local SOEs differ in their strategic attention to digitalization, a major national strategy of the government in China in recent years. The ABV proposes how various features of attention structure in organizations can jointly regulate and channel managerial attention (Barreto & Patient, Reference Barreto and Patient2013; Fu, Tang, & Chen, Reference Fu, Tang and Chen2020). We extend the original analysis of attention structure (Ocasio, Reference Ocasio1997) to the context of SOEs. One of the most important players for SOEs, the Chinese state, designs the foundational feature of attention structure, the rules of the game, and two other important features – governance structures and resources (Joseph & Ocasio, Reference Joseph and Ocasio2012; Ocasio, Reference Ocasio1997). The rules of the game are embodied in SOEs' goals, identities, and evaluation criteria of top executives. We propose how the different attention structures in central versus local SOEs channel central SOEs to allocate higher value to issues of digitalization and pay more strategic attention than local SOEs. Moreover, we propose that the specific features of governance structures and resources – pyramid structure and government subsidies – interact with the rules of the game to regulate decision-makers' attention. Given the game rule difference – primacy of national policy implementation in central SOEs versus the balanced goals in local SOEs, the pyramid structure weakens attention to digitalization in local SOEs but not in central SOEs, and government subsidies channel attention to digitalization in central SOEs but not in local SOEs.
This study investigates SOEs' focus on digitalization because, in China, such a change initiative has been mainly driven by the Chinese government. The government made digitalization a national strategy and called upon companies to fulfill this plan. However, digitalization, especially at the core of the business, can be highly risky, and not all firms see the immediate need to take such a risk without strong competitive or survival pressures (Girod & Whittington, Reference Girod and Whittington2017; Liu, Yan, Zhang, & Lin, Reference Liu, Yan, Zhang and Lin2021). Moreover, digitalization is a long-term strategy that takes a long time to bear fruit and may have an adverse impact on a firm's short-term performance (Liu et al., Reference Liu, Yan, Zhang and Lin2021). Considering that top leaders of SOEs need to show their achievements during their tenures of office, usually three to five years, they are more inclined to invest in projects with shorter income cycles and less risk (Wu, Reference Wu2012). Thus, SOEs are generally less likely to take the initiative to engage in digitalization, and the pressure from the state becomes a significant driving force. Therefore, digitalization offers an ideal context to observe whether and how central versus local SOEs attend differently to the government's national strategy. We test the arguments using all publicly listed state-owned manufacturing firms in China between 2009 and 2020 and find empirical support.
This study contributes to research on how the state influences policy implementation by understanding the differential responses of central versus local SOEs to national policy. While the literature on strong state and state capacity has focused on the role of the state in policy formulation and implementation for economic growth (Evans & Rauch, Reference Evans and Rauch1999; Guillén & Capron, Reference Guillén and Capron2016; Skocpol, Reference Skocpol, Evans, Rueschemeyer and Skocpol1985), it has not considered the structural reasons for heterogeneous national policy implementation. Local SOEs attend less to the central government pressures despite receiving government resources, especially when they are located at the far end of the control chain. In addition, this study extends the ABV by understanding the important characteristics of attention structure in different types of SOEs and how they regulate attention. Lastly, this study contributes to research on corporate digital transformation by revealing the important role of the government. The state has been promoting digitalization in many countries globally, but this burgeoning research has mainly focused on market competition and firms' internal factors to explain adoption. Our study is among the first to demonstrate how the state has channeled firms' strategic attention to digitalization in SOEs.
Theoretical Background
Attention-Based View
Ocasio (Reference Ocasio1997) proposed the ABV to understand what drives corporate decision-making. He defines attention as ‘the noticing, encoding, interpreting and focusing of time and effort by organizational decision makers’. The ABV emphasizes that the focus of decision makers' attention affects the strategic choice and outcomes of the firm (Cho & Hambrick, Reference Cho and Hambrick2006). Generally, decision-makers face multiple goals (Mair & Martí, Reference Mair and Martí2006; Ocasio, Reference Ocasio2011; Stevens, Moray, Bruneel, & Clarysse, Reference Stevens, Moray, Bruneel and Clarysse2015), but the limited rationality of decision-makers prevents them from paying sufficient attention to each goal (Cyert & March, Reference Cyert and March1963; Kahneman, Treisman, & Gibbs, Reference Kahneman, Treisman and Gibbs1992; Simon, Reference Simon1947). Thus, they selectively pay attention to the information they believe is relevant to success (Fu et al., Reference Fu, Tang and Chen2020; March & Simon, Reference March and Simon1958; Ocasio, Reference Ocasio1997; White, Reference White1992). Given competing demands for their attention, how they select issues and formulate answers is thus a key question ABV attempts to explain (Ocasio, Reference Ocasio2011).
A central tenet of the ABV is that attention is situated, that is, based on the context or situation (Nicolini & Korica, Reference Nicolini and Korica2021; Ocasio, Reference Ocasio2011). The attention structure in organizations can regulate and distribute attention. In his original framework, Ocasio (Reference Ocasio1997) suggests four important characteristics of attention structure: rules of the game, players, structural positions, and resources. The rules of the game are ‘the formal and informal principles of action, interaction, and interpretation that guide and constrain decision-makers in accomplishing the firm's tasks and in obtaining social status, credits, and rewards in this process’ (Ocasio, Reference Ocasio1997: 196). Players include the most important ones, such as top decision-makers and others who can influence the decision-making process. Structural positions refer to the system of hierarchical authority in organizations, which involves delegation of power and discretion and can shape differentiated attention to different aspects of the organizational environment. Resources are the ‘tangible and intangile assets utilized in the construction of organization moves’ (Ocasio, Reference Ocasio1997: 198), and can affect the extent certain issues are attended to and certain answers are chosen. Subsequent studies describe how resources can provide both information and incentives to motivate or hinder the strategic attention of firms (Joseph & Ocasio, Reference Joseph and Ocasio2012).
Out of the four characteristics of the attention structure, the rules of the game play a central role in distributing and channeling attention. They provide the ‘logic of action’ in assigning value and legitimacy to issues and ordering their importance. Ocasio (Reference Ocasio1997: 199) also argues for the interdependence of these characteristics in structuring attention ‘while the rules of the game are central, the allocation of the firm's legitimacy, value, is a joint product of the rules, positions, players, and resources’.
Attention Structures in Central and Local SOEs
This study extends the ABV to understand how central versus local SOEs differ in attention structures and how such differences shape their strategic attention accordingly. While SOEs are designed as tools of government intervention in the economy (Huang & Yu, Reference Huang and Yu2006), central and local SOEs diverge in some important characteristics, which lead to distinct attention structures in these two types of firms. As one of the key players in SOEs, the state designed the three characteristics of attention structures: rules of the game, structural positions, and resources.
The government administration has continued to impose on corporate decision-making in SOEs (政企不分), despite the market transition and mixed ownership reform allowing them more discretion (Chen, Lu, Liu, & Guo, Reference Chen, Lu, Liu and Guo2000; Chinese Government, 2015). Chinese SOEs are affiliated with the State-owned Assets Supervision and Administration Commission (SASAC), an important government agency that manages the country's assets. SASAC sets the goals, provides resources, and appoints and evaluates top leaders for SOEs (Simon, Reference Simon1997; Stan, Peng, & Bruton, Reference Stan, Peng and Bruton2014; Zhu, Qi, Belis, Lu, & Kerremans, Reference Zhu, Qi, Belis, Lu and Kerremans2019). Thus, the state has clear legally legitimated authority over SOEs (Monteiro & Adler, Reference Monteiro and Adler2022), and SOEs strive to perform the tasks the state gives to its satisfaction (Bloom & Van Reenen, Reference Bloom and Van Reenen2007).
Nevertheless, owned and managed by SASAC at different hierarchical levels of the state, central and local SOEs differ in the foundational feature of their attention structures: the rules of the game, which are embodied in the purpose, identity, and incentives of the organization (Ocasio, Reference Ocasio1997; Selznick, Reference Selznick1957). Regarding the purpose, central SOEs aim to accomplish the government's national strategy and serve as an exemplar in the state-owned sector in complying with the central government's expectations. Moreover, such a focus ensures that national strategies are not overshadowed so that they do not become a source of national revenue expansion (Liu & Li, Reference Liu and Li2012). In contrast, the goals of local SOEs are influenced by both the central and local governments. While they pursue the government's national strategy, they must also realize the local government's economic growth and employment targets.
The two types of firms also have distinct political identities (Genin, Tan, & Song, Reference Genin, Tan and Song2021; Meyer & Peng, Reference Meyer and Peng2016). Central SOEs are ‘national champions’. They face strong institutional pressures from the central government, and lack of compliance will lead to penalties on the top leaders and withdrawal of resource support (Genin et al., Reference Genin, Tan and Song2021; Meyer & Peng, Reference Meyer and Peng2016). In comparison, local SOEs are major contributors to the local economy and social stability (Bai & Xu, Reference Bai and Xu2005). Thus, they can be buffered from the central government's pressure if such compliance can distract them from realizing the demands from their local governments. This is particularly the case when the central government's national strategy conflicts with the local governments' short-term economic goals. For instance, studies show that local governments protected firms in heavy-polluting industries from the central government's environmental policies due to these firms' contribution to local tax revenues (Zhang, Reference Zhang2013).
Career-based incentives for top leaders of SOEs also differ, as reflected in their appointment and evaluation criteria. The central and local governments control the appointment and removal of top leaders of central and local SOEs (Yan & Xiao, Reference Yan and Xiao2019). The central government can highly value top leaders of central SOEs for following the government's national strategy (Yun, Li, & Hu, Reference Yun, Li and Hu2022). Compared with their counterparts in central SOEs, leaders in local SOEs may have less incentive to strictly follow the government's national strategy (Yun et al., Reference Yun, Li and Hu2022). Although local governments will also actively promote national strategy, they impose multiple demands on local SOEs, such as tax revenues, local economic growth, and employment.
In addition to the rules of the game, the structural positions of central and local SOEs can channel strategic attention. Here, we focus on an important governance structure, that is, the pyramid structure. Central SOEs are under the direct control of the SASAC of the National State Council, while local SOEs are under the direct control of the SASAC of provincial or municipal governments. Respectively, the agency of SASAC at central and local levels delegates its power through a control chain, the pyramidal structure (Fan, Wong, & Zhang, Reference Fan, Wong and Zhang2013), to manage multitudes of SOEs. The central SASAC exerts tight control over SOEs and holds them accountable for the central government's national strategy. Even a control chain does not diminish its control at the far end of the control chain. In contrast, there is a lack of centralized and unified supervision from local SASAC over local SOEs. Moreover, local governments tend to show supportive or relaxed control strategies based on their interests, such as an immediate increase in local fiscal revenues (Zhu, Chen, & Yu, Reference Zhu, Chen and Yu2006). This allows local SOEs more managerial autonomy and discretion, especially at the far end of the control chain (Zhang, Liu, & Cai, Reference Zhang, Liu and Cai2015).
Lastly, resources in the two types of SOEs serve different goals. Central SOEs are given priority in accessing natural resources, raw materials, technologies, energy, and financing (Li, Cui, & Lu, Reference Li, Cui and Lu2014). Indeed, they are guaranteed a plentiful and steady supply of resources, to the extent that they often pursue the government's national strategy without much consideration for costs (Lin, Cai, & Li, Reference Lin, Cai and Li1998; Liu & Li, Reference Liu and Li2012). In contrast, there is a reciprocal relationship between local SOEs and local governments (Chen, Kong, & Wang, Reference Chen, Kong and Wang2016; Peng, Bruton, Stan, & Huang, Reference Peng, Bruton, Stan and Huang2016). Since local governments expect local SOEs to help with local economic growth and employment, they provide SOEs with some preferential access to resources compared with non-SOEs. However, given the limited resources of local governments, local SOEs also need to acquire resources through their own efforts. Moreover, local governments usually subsidize loss-making local SOEs to help them survive market competition (Röller & Zhang, Reference Röller and Zhang2005; Wei, Varela, & Hassan, Reference Wei, Varela and Hassan2002).
In summary, the differences between central and local SOEs suggest different attention structures. Table 1 summarizes the key differences.
Different attention structures in central and local state-owned enterprises
Digitalization: SOEs' Response to the State's National Strategy
Digitalization of enterprises refers to the process of enabling production, management, marketing, and other corporate functions through digital technologies, including artificial intelligence, big data, cloud computing, and blockchain. The goal of digitalization is to help firms improve efficiency as well as transform and upgrade manufacturing (Bharadwaj, Sawy, Pavlou, & Venkatraman, Reference Bharadwaj, Sawy, Pavlou and Venkatraman2013; Bodrožić & Adler, Reference Bodrožić and Adler2021; Hanelt, Bohnsack, Marz, & Antunes, Reference Hanelt, Bohnsack, Marz and Antunes2020). In many countries, the digitalization of manufacturing firms is regarded as an important national strategy, such as ‘Industry 4.0’ in Germany and the ‘National Strategy for Advanced Manufacturing’ of the United States. In China, the central government views digitalization as a window of opportunity to narrow the manufacturing gap with developed countries. It has adopted many digital-related policies to promote the digitalization of the manufacturing industry, such as ‘Made in China 2025’ (Corrocher, Mavilia, & Giorgio, Reference Corrocher, Mavilia and Giorgio2018).
However, digitalization presents many challenges and uncertainties for firms. Digitalization, especially at the core of the manufacturing process, can be highly risky. Such a strategic change not only involves the application of digital technologies but also requires tight integration of digital technologies with all aspects of organizations (Tao, Cheng, Qi, Zhang, Zhang & Sui, Reference Tao, Cheng, Qi, Zhang, Zhang and Sui2018). As a result, the digitalization of China's manufacturing industry has remained in the exploratory stage, and no mature paradigm or success template has emerged (Girod & Whittington, Reference Girod and Whittington2017). Accenture's 2020 report shows that only 7% of firms in China have achieved significant results in digital transformation. Moreover, compared with developed countries, China only started industrialization in recent decades, and the foundation of industrialization is still relatively weak (Chu et al., Reference Chu, Kim, Jang, Song, Rodrigue, Chun, Cho, Ko, Cho, Cha, Min, Jeong, Jeong, Lee, Chu and Ahn2016). Thus, many firms cannot effectively use digital technologies, and digitalization has led to the efficiency paradox (Brynjolfsson, Rock, & Syverson, Reference Brynjolfsson, Rock and Syverson2019). Firms invest huge amounts of resources in digitalization, but do not reap efficiency gains or even lose efficiency due to low proficiency in the use of digital technologies (Autor & Murnane, Reference Autor and Murnane2003; Bresnahan, Brynjolfsson, & Hitt, Reference Bresnahan, Brynjolfsson and Hitt2002). Lastly, corporate digitalization is a long-term project that may lead to declining short-term financial performance (Liu et al., Reference Liu, Yan, Zhang and Lin2021). For example, Wang Lei, a manager at Hefei Metal Forming Intelligent Manufacturing Co., Ltd., mentioned in a First Financial interview that the company's digital investment projects have not generated income in the past five years, and it will take another two to three years to become a new profit growth point of the company. Given that SOEs' top leaders are evaluated based on short-term firm performance, they can be reluctant to engage in such a strategic change (Wu, Reference Wu2012).
Nevertheless, SOEs are expected to engage in digitalization, given their objective of national strategy implementation. Prior studies have established that SOEs are important policy tools because of the domination by superior administrative authorities and evaluation criteria based on the fulfillment of the national strategy (Zhu et al., Reference Zhu, Qi, Belis, Lu and Kerremans2019). According to this view, resources and opportunities of SOEs are mainly provided by the government departments and agencies, major corporate decisions need to be approved by superior administrative authorities, and top managers are appointed and evaluated by the Organization Department of the state (Koppell, Reference Koppell2007). State ownership and control make SOEs essential in addressing social problems and creating public value (Cappellaro, Tracey, & Greenwood, Reference Cappellaro, Tracey and Greenwood2020; Lin & Tan, Reference Lin and Tan1999). Moreover, the state's long-term strategic plan emphasizes the high quality and sustainable development of SOEs through digitalization, so that the Chinese economy can gain an edge in the international competition of digitalization. Therefore, SOEs are expected to play an important role in accomplishing the government's national strategy of digitalization and leading the development of a modern industrial system (Qi, Du, & Wen, Reference Qi, Du and Wen2021). In this context, digitalization has become an important political mission for SOEs.
Hypotheses Development
Central and Local SOEs' Different Strategic Attention on Digitalization
We propose that the different attention structures in central and local SOEs can distribute top decision-makers' attention to digitalization differently. First, different rules of the game can lead top decision-makers to assign different levels of priority and urgency to the state's agenda (Ocasio, Reference Ocasio2011). For central SOEs, the goal of fulfilling the government's national strategy can channel their attention to digitalization. Issues related to digital transformation are perceived as worthy of time and resources, given this goal's high legitimacy and priority. In addition, the identity of ‘national champions’ can guide decision-makers' attention to issues that can strengthen rather than undermine their identity (Wry & York, Reference Wry and York2017). For example, previous studies find that social enterprises tend to focus more on social issues than on how to make a profit due to their socially-oriented identity (Ebrahim, Battilana, & Mair, Reference Ebrahim, Battilana and Mair2014). Central SOEs view themselves as leaders of the entire industry (Liu, Reference Liu2021), and thus attend to the national strategy of digitalization before other firms start to do so. Furthermore, incentives used in the appointment and assessment of top leaders can regulate and direct their attention to issues most valued. Top leaders in central SOEs are appointed directly by the central government. Their career goal of political promotion is likely to focus their attention on the central government's national digitalization strategy because fulfillment of the national policy helps them obtain better evaluation from the central government (Yun et al., Reference Yun, Li and Hu2022).
In contrast, for local SOEs, the dual goals of complying with national strategies and achieving local government's economic and social objectives can dilute the local SOEs' attention to the central government's policies. In addition, local SOEs' identity as key contributors to local economic well-being and social stability can focus their attention on the responsibility of boosting local economic development, which has been emphasized by local governments (Cao, Liu, & Zhang, Reference Cao, Liu and Zhang2009; Chen et al., Reference Chen, Kong and Wang2016). Such an economic burden often precedes compliance with the government's national strategy for local SOEs (Chen et al., Reference Chen, Kong and Wang2016). Furthermore, career-based incentives for top leaders in local SOEs guide them towards issues and solutions that are more rewarded by local governments. The local government appoints these top leaders whose needs (such as short-term local revenues and employment) are much broader than the national digitalization strategy. The evaluation of top leaders of local SOEs is not as tightly based on the national strategy implementation as it is on central SOEs. The broader set of goals – both national and local government objectives – linked to top leaders' evaluation in local SOEs provides them with more managerial autonomy and reduces their focus on digitalization.
Second, the governance structures in the two types of SOEs also influence the distribution of top decision-makers' attention. As central SOEs are under the direct and tight control of central SASAC (a subsidiary of China's State Council), reporting procedures, communication channels, and monitoring systems can focus top leaders' attention on the state objective of digitalization. In contrast, local SOEs report to the local SASAC (an agency of local government), and local SASACs do not exert a unified force of control (Zhang et al., Reference Zhang, Liu and Cai2015). This gives local SOEs' top leaders greater discretion and leads them to attend to the diverse set of goals of local government.
Third, resources often bear information that can influence firms' attention (Hu, Hughes, & Hughes, Reference Hu, Hughes and Hughes2022; Joseph & Ocasio, Reference Joseph and Ocasio2012). Digitalization involves high investment and uncertain outcomes (Liu et al., Reference Liu, Yan, Zhang and Lin2021), and thus requires firms to have sufficient resources to implement such strategic transformation. For central SOEs, the abundance of resources provided by the central government conveys the information about its full support for the risky implementation of digitalization (Li et al., Reference Li, Cui and Lu2014). However, despite local SOEs' better access to local government-controlled resources compared with private firms, local SOEs still need to obtain resources through their own efforts. The resources transmit the information that local SOEs need to achieve local economic growth as well as national strategic objectives, and hence local SOEs are less likely to focus their strategic attention on digitalization.
In short, due to their different attention structures, central SOEs are more likely to give more strategic attention to digitalization than local SOEs. Thus, we propose:
Hypothesis 1 (H1): Central SOEs exhibit a stronger strategic attention to digitalization than local SOEs.
Variation in Central versus Local SOEs' Strategic Attention on Digitalization
The attention structures among central SOEs can still differ substantially, especially concerning the characteristics of governance structures and resources, while the same holds among local SOEs. The SASAC adopts the pyramid structure to manage multiple SOEs (Fan et al., Reference Fan, Wong and Zhang2013). Those firms at the far end of the control chain are situated in different structural positions than those at the close end, and their attention can be shaped differently. In addition, firms are provided with different amounts of government subsidies, and such resources can provide information and incentives to channel firm attention toward the state's objectives accordingly. More importantly, the rules of the game in attention structure play a foundational role in providing top leaders with the schema of decision-making, and how governance structures and resources regulate attention can be influenced by the rules of the game (Ocasio, Reference Ocasio1997). We propose how the specific features of governance structures and resources – pyramid structure and government subsidies – can influence attention focus differently in central versus local SOEs, given their different rules of the game.
Governance structure – Chain of control through pyramidal structure
Joseph and Ocasio (Reference Joseph and Ocasio2012) argue how cross-level organizational structural linkages can influence the structural distribution of attention. The pyramid structure affects the involvement and influence of the ultimate firm controller. Firms structurally placed at several levels below the center of power can have more discretion in their decision-making, and their attention is less influenced by the ultimate controller (Belenzon, Hashai, & Patacconi, Reference Belenzon, Hashai and Patacconi2019). While SOEs are under the administrative control of SASAC, government officials from this agency do not intervene in multitudes of firms to the same extent. The agency relies on the chain of control to achieve delegation of power through an important governance mechanism of the pyramidal structure (Fan et al., Reference Fan, Wong and Zhang2013). As shown in Figure 1, the SASAC of Guangdong Province controls the listed local SOE, Shenzhen Zhongjin Lingnan Non-ferrous Metal Co., Ltd. (the focal firm), through a pyramid structure. Specifically, Guangdong SASAC owns Guangdong Guangsheng Asset Management Co., Ltd. 100% (called Guangsheng). Guangsheng controls the focal firm through three control chains. It directly controls 27.51% of the focal firm and indirectly controls 4.33% (1*0.8862*0.0391 + 1*0.0086) of the focal firm. The length of the control chain is 3 as ‘Guangdong SASAC-Guangsheng-focal firm’ is the main control chain.
Pyramid structure of Shenzhen Zhongjin Lingnan Non-ferrous Metal Co., Ltd.
When the power of the SASAC is largely delegated, the state power is decentralized. Communication of state policy may not be timely, and monitoring of policy implementation may be weakened, resulting in a reduced policy burden on these firms at the far end of the control chain (Zhang, Lijun, Zhang, & Yi, Reference Zhang, Lijun, Zhang and Yi2016). Meanwhile, as it is difficult for the government to gather specific information about these SOEs, the cost of government intervention in their daily operation also increases (Chen, Sun, Tang, & Wu, Reference Chen, Sun, Tang and Wu2011). Some scholars argue that a high level of delegation can effectively improve the professionalism of management, efficiency, total factor productivity, and profitability of SOEs (Cao, Xu, Lu, & Tang, Reference Cao, Xu, Lu and Tang2015; Fan et al., Reference Fan, Wong and Zhang2013). As a result, when the distance between the SOE and the SASAC becomes larger, the SOE can pay less attention to the state's national objectives.
However, the pyramid structure can affect strategic attention differently in central and local SOEs, given their different rules of the game. The overarching goal and organizational identity of fulfilling national objectives in central SOEs can constrain decision-makers' attention to digitalization. The central government exerts its control through multiple channels, such as evaluation based on national policy implementation for top leaders. In comparison, in local SOEs, as their objective and identity of helping local governments have channeled their attention to multiple goals, power delegation can further weaken the institutional pressure and in turn compromise firms' attention to national policy (Jiang, Reference Jiang2016). Thus, an extended control chain of the pyramid structure weakens local SOEs' attention to the government's national strategy, such as digitization, more than it does for central SOEs. Hence, we propose:
Hypothesis 2 (H2): The chain of control through pyramidal structure decreases local SOEs' strategic attention to digitalization more than it does that of central SOEs.
Government subsidies
Ocasio (Reference Ocasio1997) suggests that resources can be utilized to serve different goals depending on the logic of action in an organization and that resources alone cannot determine attention. Resources transmit the will of resource providers (owners or those with access) and channel firms' strategic attention. For example, because family business owners often prioritize preserving family wealth, their resources divert family businesses' strategic attention away from risky activities, such as radical innovation (Hu et al., Reference Hu, Hughes and Hughes2022). Resource allocation is essential for the state to ‘guarantee an effective implementation of policies’ (Weaver & Rockman, Reference Weaver and Rockman1993: 446). Government subsidies are an important type of government resource, transmitting information about the state's expectations and providing incentives to achieve the state objectives.
We argue that government subsidies affect the distribution of attention to digitalization differently in central and local SOEs because the rules of the game can render the subsidies to be used to serve different goals (Stan et al., Reference Stan, Peng and Bruton2014). As central SOEs are ‘national champions’ expected to fulfill national strategy (Li et al., Reference Li, Cui and Lu2014; Wang, Hong, Kafouros, & Wright, Reference Wang, Hong, Kafouros and Wright2012), government subsidies are to be used to, first and foremost, support their compliance with the government's national strategy. In contrast, local SOEs' goals of fulfilling the economic growth and employment targets from local governments can constrain the use of government subsidies. The failure of local SOEs will cause a blow to the local economy and employment, which will have a negative impact on the local government (Peng et al., Reference Peng, Bruton, Stan and Huang2016). Moreover, there is a reciprocal relationship between local SOEs and local governments (Chen et al., Reference Chen, Kong and Wang2016; Peng et al., Reference Peng, Bruton, Stan and Huang2016). Generally, government subsidies are provided to help local SOEs survive market competition, which in turn can help the government gain more resources and revenues (Liu & Li, Reference Liu and Li2012; Röller & Zhang, Reference Röller and Zhang2005; Wei et al., Reference Wei, Varela and Hassan2002).
Therefore, when a higher level of subsidies is provided, more information is conveyed to the central SOEs than to local SOEs about the state's objectives related to digitalization, and more incentives are given to central SOEs to exert their efforts on the national digitalization strategy at all costs. Thus, government subsidies can make the national strategy even more salient in central SOEs, channeling their attention toward digitalization. Hence, we propose:
Hypothesis 3 (H3): Government subsidies increase central SOEs' strategic attention to digitalization more than they do that of local SOEs.
Methods
Samples and Data
This study examines how central and local SOEs differ in their strategic attention to digitalization in China. A database of Chinese-listed manufacturing SOEs from 2009 and 2020 was assembled to test the hypotheses. The manufacturing industry was chosen as the sample because the digital transformation of the manufacturing industry is one of the top concerns for the Chinese government, and the digitalization of the manufacturing industry is less developed than that of other industries, such as the service industry, in China.
Data were collected on firm performance, firm characteristics, and demographic characteristics of top leaders from the China Stock Market & Accounting Research Database (CSMAR), which is one of the most comprehensive databases for listed companies in mainland China and has been used in previous studies (Chen, Tang, Wu, & Yang, Reference Chen, Tang, Wu and Yang2021; Giannetti, Liao, You, & Yu, Reference Giannetti, Liao, You and Yu2020). Texts of the management discussion and analysis (MD&A) in annual reports from the Chinese Research Data Services Platform (CNRDS) were obtained. Provincial indicators were gathered from the National Bureau of Statistics of China. Data on digitalization policies was obtained from PKULAW, which collects policies at all levels in China, containing the title, content, and issuing agency of policies, and has been used in prior research (Nan, Feng, Hu, & Qi, Reference Nan, Feng, Hu and Qi2020; Yang & Huang, Reference Yang and Huang2022). After merging the databases and excluding missing information, we obtained 4,495 firm-year observations on 545 unique firms. Among them are 365 local SOEs with 2,783 firm-year observations, accounting for 61.97% of the full sample.
Dependent Variable
Strategic attention on digitalization
Following prior studies (Yu, Wang, & Li, Reference Yu, Wang and Li2020; Zhao, Reference Zhao2021), we used text analysis to evaluate the strategic attention on digitalization. The annual report's management discussion and analysis (MD&A) is used for text analysis. MD&A is a chapter that must be disclosed in the annual report of a listed company in China. Its content includes the operation and management of the enterprise, the development of the industry, the analysis of policies, and the deliberation and planning of the enterprise's future development strategy. Therefore, MD&A reflects a firm's strategic attention, and coding based on its text analysis is a valid measure (Wang, Zhang, & Shang, Reference Wang, Zhang and Shang2022).
To analyze the text of MD&A, a corpus of keywords related to digitalization was first constructed. Python was used to conduct word segmentation and compute word frequency statistics for the MD&A in all the annual reports. Two coders separately reviewed the results of word segmentation and identified the keywords related to digitalization, which became the initial corpus. A high interrater reliability of 0.89 was obtained, indicating the initial corpus's high reliability (Short, Broberg, Cogliser, & Brigham, Reference Short, Broberg, Cogliser and Brigham2010). A word segmentation was then conducted for academic papers on digitalization in authoritative Chinese academic journals and keywords related to digitalization. These keywords represent the academic description of corporate digitalization. The initial corpus was expanded with the non-overlapping keywords extracted from the academic papers. Finally, six experts from government and digital associations with rich experience in digitalization were invited to evaluate the corpus and help improve it. The final corpus shown in Appendix I describes digital technologies and applications such as big data, cloud computing, artificial intelligence, and blockchain.
Then, MD&A was retrieved based on the corpus to identify the sentences containing keywords related to digitalization. The strategic attention of SOEs on digitalization was measured by calculating the proportion of sentences describing digitalization in the MD&A (Levy, Reference Levy2005). Figure 2 shows the trend of strategic attention on the digitalization of SOEs over time. Central SOEs pay more strategic attention to digitalization than local SOEs.
Trend of strategic attention on digitalization in SOEs
Independent Variable
Central SOE
Whether an SOE is local or central depends on the hierarchy of the ultimate controller. Central SOE is a dummy variable coded 1 if the ultimate controller is the central government (specifically, controlled by the SASAC of the National State Council), and 0 if the ultimate controller is the local government (controlled by the SASAC of the local government, which refers to either provincial or city level local government).
Moderating Variables
Chain of control through the pyramidal structure
The pyramid structure of firm ownership was used to measure an SOE's position in the chain of control. Previous research suggests that a long control chain is associated with a high degree of government power delegation (Fama & Jensen, Reference Fama and Jensen1983; Fan et al., Reference Fan, Wong and Zhang2013; Jiang, Reference Jiang2016). Length of Control Chain was used to measure the number of intermediate layers between the listed company and its ultimate controller (i.e., SASAC at the central or local level, respectively) (Zhang et al., Reference Zhang, Lijun, Zhang and Yi2016). For example, the Length of Control Chain takes the value 1 if the ultimate controller directly controls the listed company, and the value 2 if there is an intermediate controller between the ultimate controller and the listed company, and so on. Alternatively, the number of control chains (i.e., the number of different chains through which the ultimate controller can exercise control over the listed company) was also used, which some studies suggest also indicates delegation and decentralization of state power (Su, Reference Su2012). The two variables were calculated based on the chart of shareholder control relationships in corporate annual reports.
Government subsidies
The value of subsidies that firms received from the government each year was obtained through CSMAR. For missing values, the information was collected manually from corporate annual reports. Listed firms are required to disclose the subsidies they receive from the government (Lee, Walker, & Zeng, Reference Lee, Walker and Zeng2014). Government Subsidies is measured as a percentage of total assets to adjust for the influence of company size.
Control Variables
Firm characteristics that prior studies suggest are important for shaping strategic attention on digitalization were controlled (Eller, Alford, Kallmünzer, & Peters, Reference Eller, Alford, Kallmünzer and Peters2020). Firm age can make firms less adaptable to new technology. Firm size is measured as number of employees (logged), and firm performance as ROA. Leverage is measured as the asset–liability ratio. Ownership by non-state shareholders (versus state ownership) in these listed firms can affect the extent of government influence (Maw, Reference Maw2002). The nature of the top ten shareholders of these SOEs was examined, and the shareholding ratio of state-owned shareholders (Ma, Wang, & Zhang, Reference Ma, Wang and Zhang2015) was calculated as State-owned shares. Firms with more innovative capability may pay more attention to digitalization and the proportion of employees with college degrees or above (high-education employees) was controlled. Moreover, slack resources can help firms experiment with strategic transformation (Greenley & Oktemgil, Reference Greenley and Oktemgil1998), and we controlled firms' slack resources as the ratio of net cash flow from operating activities to total assets.
We controlled for top leaders' characteristics as important decisions in SOEs are made through the joint consultation of the board members and top management team members. The average age of these leaders was controlled. Younger leadership teams tend to be more familiar with digital technology and more willing to pay attention to digitalization. We also controlled their average education, calculated as the percentage of board and top management team members with graduate degrees or above. Finally, we controlled their work experience, measured as the total number of different departments in which these top leaders have worked.
We also controlled for the characteristics of the province where SOEs are headquartered. Local SOEs contribute fiscal revenues to local government, which can be instrumental in filling the government's fiscal gap (Zhao & Yang, Reference Zhao and Yang2015). Therefore, when the local economy is in a downturn, local SOEs in particular may face greater economic pressure. Hence, we controlled Provincial GDP Growth Rate. Moreover, the extent of government intervention in the market varies in different regions, which can affect firms' response to the government's national strategy. Thus, we controlled the degree of marketization in each province (Wang, Hu, & Fan, Reference Wang, Hu and Fan2021).
Firms tend to imitate their peers, which generally refer to firms in the same industry, as they face similar technical and institutional environments. A similar technical environment means that firms face similar demands from suppliers, customers, and technological trends, and hence, their strategic attention can share commonalities (Allee, Do, & Sterin, Reference Allee, Do and Sterin2021; Houston, Lev, & Tucker, Reference Houston, Lev and Tucker2008). Two measures were created accordingly given the difference between SOEs and non-SOEs (e.g., non-SOEs are more market-oriented). SOEs' average digital attention is measured by the average of the listed SOEs' digital attention in each industry (excluding the focal firm itself). Non-SOEs' average digital attention is measured by the average of the listed non-SOEs' digital attention (measured similarly as described in the dependent variable above) in each industry. These two measures were then orthogonalized to reduce collinearity. The industry of a firm was delineated according to the industry classification standard issued in China (GBT4754-2017). We also controlled the industry competition measured by the Herfindahl index, which reflects the concentration of the industry.
Finally, the number of national digitalization policies and provincial digitalization policies issued each year (logged) were used to represent the extent to which the government promotes corporate digitalization. We used the keywords retrieval method to evaluate whether a policy is related to digitalization. The keywords included intelligent, intelligent manufacturing, industrial internet, internet plus, automation, big data, 5G, robot, internet, IOT, cloud, and blockchain. These keywords were chosen because they appeared in the most representative Chinese government documents, such as China's five-year Plan and Made in China 2025. Descriptions in other Chinese policies were generally consistent with these programmatic documents. National digitalization policies and local digitalization policies were orthogonalized to reduce collinearity.
Finally, dummy variables were included for each year to take the temporal effect into account. Province dummies were also included in the model. All the independent and control variables are measured at t, and the dependent variable is measured at t + 1.
Results
Table 2 presents the descriptive statistics and correlations for the variables. The correlation coefficients of all the variables are between −0.44 and 0.39. Hence, multicollinearity is unlikely to be a problem for our analysis. Strategic attention to digitalization is positively related to Central SOE (r = 0.11, p < 0.001). Variance inflation factor tests were also performed. The results showed that the variance inflation factors for all variables across all models were below 2, far below the empirical threshold of 10, confirming that multicollinearity was less likely for analysis.
Correlations and descriptive statistics
Notes: *p < 0.05; **p < 0.01; ***p < 0.001 (two-tailed test).
Since the independent variable, central SOE, is a dummy variable that hardly changes over time, the random effects models were used. In Table 2, Column (1) contains only the control variables. Column (2) includes the independent variable, Central SOE. The coefficient of Central SOE is 0.810 (p ≤ 0.05), meaning that the proportion of sentences related to digitalization in the MD&A in central SOEs is 0.810% higher than in local SOEs. Such a difference is substantial, given that the average proportion of digital-related discussion in the full sample is only 2.52%. Hence, central SOEs pay more attention to digitalization than local SOEs. Thus, H1 is supported.
To test H2 and H3, interaction terms were included in Columns (3) and (4), respectively. Model 5 includes all the interaction terms and is the full model. Continuous variables in the interaction terms were standardized.
In Column (5), the interaction term of Central SOE * Length of Control Chain (β = 0.663, p < 0.05) is positive, indicating that the length of control chains moderates the relationship between central SOE and strategic attention on digitalization. All continuous variables were centered before generating interaction terms. Thus, the two-way interaction is the moderation effect of the length of the control chain on central SOEs at its mean, which means that at the average level of control chain length, the digital attention of central SOEs is 0.633% higher than that of local SOEs. Figure 3 is also plotted to illustrate the interaction effects. With the short control chain, the difference between central SOEs and local SOEs regarding their strategic attention to digitalization is small. However, with the long control chain, the strategic attention on digitalization in central SOEs is significantly higher than that in local SOEs. This suggests that power delegation significantly reduces local SOEs' attention to the state's national strategy but has little impact on that of central SOEs. These results lend support to H2.
Moderating effect of length of control chain on the relationship between central SOE and strategic attention on digitalization
In Column (4), the interaction term of Central SOE * Government subsidies is positive (β = 1.249, p < 0.05), which means that at the average level of subsidies, the digital attention of central SOEs is 1.249% higher than that of local SOEs. As shown in Figure 4, compared with the line for low-level government subsidies, the slope of the line for high-level subsidies is steeper. When the level of subsidies is high, the difference between the two types of SOEs is further enlarged. Thus, Hypothesis 3 is supported.
Moderating effect of government subsidies on the relationship between central SOE and strategic attention on digitalization
Regarding control variables, both average digital attention of SOEs and non-SOEs in the Industry have a significant positive effect on the focal SOE's strategic attention on digitalization. However, the effect of average digital attention of SOEs is significantly stronger than that of non-SOEs (p = 0.000). Firm size also has a significant positive effect. This is probably because large SOEs are more visible targets for institutional pressures (e.g., Luo et al., Reference Luo, Wang and Zhang2017). Moreover, a higher proportion of employees with college degrees or above attend more to digitalization, suggesting that a higher level of knowledge and innovation capability enables firms to better integrate digital technology (Bresnahan et al., Reference Bresnahan, Brynjolfsson and Hitt2002). Moreover, the average age of the leadership team has a significant negative effect, which indicates that younger teams pay more attention to digital transformation. Finally, both national digitalization and provincial digitalization policies have a significant positive effect, which is consistent with the assumption that the government plays an important role in promoting SOEs' digitization.
Robustness Test
Alternative measures of a firm's strategic attention to digitalization were used. First, we used the number of sentences containing descriptions related to digitalization (logged) instead of the ratio. The hypotheses remained supported. Second, we calculated the dependent variable according to the digital keywords that Zhao (Reference Zhao2021) disclosed and similar results were obtained. Third, we tried a non-text-based measure of digitalization – investment in non-tangible digital assets. These data were obtained from the non-tangible assets table of firms' annual reports according to Zhang, Li, and Xin (Reference Zhang, Li and Xin2021). If the items of non-tangible assets included keywords related to digital transformation technologies, they were identified as non-tangible digital assets. The annual increment of non-tangible digital assets (logged) was calculated as the dependent variable. H1 and H2 were supported, while the interaction for H3 was in the same direction but non-significant.
Moreover, the impact threshold of a confounding variable (ITCV) was also estimated following Busenbark, Yoon, Gamache and Withers (Reference Busenbark, Yoon, Gamache and Withers2021). This threshold test aims to assess how sensitive the results are to potentially omitted variables. The threshold for our model was 0.211. Only the partial correlation between ‘SOEs' Average Digital Attention’ and the dependent variable in our model exceeded this threshold. We think this result was acceptable because it is difficult to find a variable as relevant as ‘SOEs' Average Digital Attention’ to the dependent variable. Even the partial correlation between ‘non-SOEs' Average Digital Attention’ and the dependent variable was less than this threshold.
Discussion
Our study is motivated by the lack of understanding of how different types of SOEs respond to national policies differently. Drawing on the ABV, we argue that different attention structures in central and local SOEs can shape top leaders' strategic attention accordingly. In the context of SOEs' response to the state's national strategy for digitalization in China, we found evidence consistent with our argument. The results show that central SOEs pay more strategic attention to digitalization than local SOEs. Moreover, the chain of control decreases local SOEs' attention to digitalization more than it does that of central SOEs. Central SOEs pay even more attention to digitalization than local SOEs when these central SOEs receive more government subsidies. These results suggest how the characteristics of attention structures, namely, governance structures and resources, can be influenced by the game rules (which differ in central versus local SOEs) in shaping strategic attention.
This study contributes to a deeper understanding of the heterogeneity of SOEs. Most studies emphasize the difference between SOEs and non-SOEs in their compliance with national policies (Lin & Li, Reference Lin and Li2008; Lin & Tan, Reference Lin and Tan1999). Some studies that examine the difference between central and local SOEs tend to focus on how the different objectives of central versus local governments influence firms owned by the different levels of government (Luo et al., Reference Luo, Wang and Zhang2017). This study systematically develops a framework to understand the differences in attention structures between central and local SOEs and demonstrates how these differences shape their strategic attention to national policy. By understanding different responses to digitalization, this research discovers a consequence when the national policy is not necessarily aligned with the short-term objectives of local governments: local SOEs pay less attention to the national digital transformation strategy than central SOEs. Our study reveals how central SOEs are important for effective policy implementation, as well as a less-noted structural reason why national policy implementation can be compromised – local SOEs' attention structure characteristics.
Our study also extends the ABV to the context of SOEs. First, based on Ocasio's (Reference Ocasio1997) original framework of the important characteristics of attention structure, it depicts how two types of SOEs differ in these characteristics. The findings of this study reinforce the core tenet of ABV, which is that attention is situated, by showing how attention to national strategy diverges among SOEs due to their different attention structures. Second, although ABV recognizes the interdependence of the characteristics of attention structure, our study advances the understanding of how governance structure and resources can regulate decision-makers' attention differently, given the different rules of the game. While previous studies focus on how the governance structure of the control chain provides more managerial autonomy (Fan et al., Reference Fan, Wong and Zhang2013), this study demonstrates that attention is distributed differently at the far end of the control chain between central SOEs and local SOEs. In other words, the length of the control chain will not reduce attention to the state policy from the top if the organization's goal, identity, and incentives are tightly centered on carrying out the state policy. Recognizing the interplay between governance structure and game rules can better explain the structural distribution of attention. Similarly, while prior studies have examined the role of government subsidies in SOEs, such as increasing R&D investments and reducing innovation efficiency (Zhou, Gao, & Zhao, Reference Zhou, Gao and Zhao2017), this study uncovers the different effects of government subsidies on SOEs' strategic attention to national policy in central versus local SOEs.
Lastly, our study also contributes to research on corporate digitalization. Most of this literature has focused on the market and the internal characteristics of organizations that facilitate or impede this strategic change (Bharadwaj et al., Reference Bharadwaj, Sawy, Pavlou and Venkatraman2013; Bodrožić & Adler, Reference Bodrožić and Adler2021; Hanelt et al., Reference Hanelt, Bohnsack, Marz and Antunes2020). However, given the high-risk and long-term nature of digital projects in manufacturing, other driving forces, such as the state, may be critical. The study adopts the perspective of digitalization as firms' responses to the government's national strategy and underscores the state's role in different types of SOEs. Our framework helps to better explain SOEs' digital decisions in contexts where the government's national strategy largely drives digitalization. Given the global state policies that promote digital transformation, it may also have implications for other countries.
Limitations and Future Research Directions
Nevertheless, this study has some limitations. First, in the theoretical framework on the attention structure, the focus is only on one important player – the state, which shapes the rules of the game, structural position, and resources for SOEs. However, according to Ocasio (Reference Ocasio1997), players that affect firms' attention are diverse, including internal and external actors. Future research can further examine how other actors, such as private shareholders or different levels of management, can shape SOEs' attention. Second, due to data availability, the sample includes only listed SOEs. Non-listed SOEs can also exhibit the pattern argued, and future studies can examine them to test our arguments further. Third, the digital strategic attention of SOEs was measured through text analysis of MD&A and investment in non-tangible digital assets. While these variables reveal managerial attention well, different measures, such as questionnaires, other corporate efforts to digitalize, and resources spent on digitalization, can illustrate strategic attention in a well-rounded manner. Fourth, this study did not address how differences in digital attention of central and local SOEs can change over time. When market competition based on digitalized manufacturing intensifies, the market forces may affect these two types of firms differently. A significant change in the difference over time was not observed in the limited time window. Future research can examine a longer period and explore such temporal change. Finally, this research investigates differences between central and local SOEs in only one country. Future research can examine this issue in different political regimes and explore the generalizability of the arguments.
In sum, our study extends the ABV to understand the structural sources of heterogeneity in central versus local SOEs regarding their responses to national policy. It reveals the distinct and varied attention structures in SOEs and the impact of such structures on their digital transformation.
Random-effect regression models predicting central and local SOEs' strategic attention on digitalization in China, 2009–2020
Notes: Robust standard errors in parentheses. †p < 0.1, *p ≤ 0.05, **p < 0.01, ***p < 0.001 (two-tailed test).
Acknowledgements
We sincerely thank the Guest Editor, Professor Dr. David Zhu, and the three anonymous reviewers for their insightful suggestions and contributions. We also appreciate Professor Dr. Kenneth G. Huang for his valuable suggestions in the paper developing workshop. We acknowledge the National Natural Science Foundation of China (Grant No. 72072020), and the Sichuan Philosophy and Social Sciences Foundation (Grant No. SCJJ24ZD23) for financial support.
Data availability statement
The data that support the findings of this study are openly available in the Open Science Framework at https://osf.io/n9e8x/?view_only=049e8322a78546809ad1199200eba15f
The keyword list of digitalization (Translated from the Chinese keywords list)
Xiaowei Rose Luo (xiaowei.luo@insead.edu) is a tenured senior faculty member in the area of Entrepreneurship and Family Business at INSEAD. She is a prominent scholar in the fields of institutional theory, corporate political strategy, entrepreneurship, emerging market strategies, and family business research. She has published numerous research studies in leading academic journals such as Administrative Science Quarterly, Organization Science, Academy of Management Journal, and Strategic Management Journal, and serves as editorial board members in leading journals.
Jun Wang (202021150311@std.uestc.edu.cn) is a PhD candidate at the School of Management and Economics, University of Electronic Science and Technology of China. His research interests encompass the governance of state-owned enterprises, along with innovation and transformation strategies. He also pays close attention to the changes and potential influences of Chinese enterprises' innovation internationalization strategies against the backdrop of deglobalization.
Shuangying Chen (shychen@uestc.edu.cn) is a professor at the School of Management and Economic, University of Electronic Science and Technology of China. Her research interests are mainly in the field of digitalization and innovation strategies in emerging market, with a particular focus on family firms and state-owned enterprises in China. She has successfully gained several competitive research funds from NSFC and Ministry of Education Humanities, etc.
Boyi Chen (lonnin@163.com) is a PhD candidate at the School of Management and Economics, University of Electronic Science and Technology of China. Her research focuses on innovation and transformation in family businesses and state-owned enterprises, with a particular interest in intergenerational succession and strategic change in family firms and digital transformation of state-owned enterprises in emerging market.
Open access
