Innovation in Manufacturing Within the Digital Intelligence Context: Examining Faultlines Through Information Processing

Abstract

In the context of digital intelligence, innovation is vital for manufacturing enterprises to establish sustainable competitive advantages. As the cornerstone of decision-making, the information-processing capability of top management teams plays an essential role in driving organizational success. Using panel data from A-Share manufacturing listed companies between 2015 and 2023, we conducted programming in the R language employing hierarchical clustering and k-means algorithms for faultline grouping calculations. The empirical analysis portion utilized STATA software, where the Hausman test was implemented to determine the use of a fixed-effects model for computation. The results demonstrate that task-related faultlines, driven by factors such as educational background, tenure, career experience, and years of service, have a positive impact on innovation performance. In contrast, relationship-related faultlines influenced by gender and age exhibit a negative effect. Furthermore, long-term investment decision preferences mediate the relationship between faultlines and innovation performance. Performance expectation gaps amplify the positive influence of task-related faultlines and mitigate the negative effects of relationship-related faultlines. In comparison with the majority subgroup, when the chairperson is part of a minority subgroup, the faultline has a more significant impact on innovation performance. This study presents a novel framework for fostering innovation within the manufacturing industry under the digital intelligence context. By combining R programming with empirical analysis, we thoroughly examine how the characteristics of top management teams’ faultlines influence innovation performance through an information processing perspective. Our findings provide actionable insights for optimizing executive structures and aligning decision-making strategies, thereby advancing organizational effectiveness.

1. Introduction

Under the backdrop of digital intelligence, innovation activities in manufacturing enterprises increasingly depend on the extraction and analysis of vast amounts of information. In the process of acquiring, integrating, and analyzing information, the top management team, as the core of strategic decision-making within organizations, plays a crucial role. Manufacturing enterprises seeking to enhance innovation efficiency must construct a top management team proficient in maximizing the utilization of information and making strategic decisions that promote innovative outcomes. According to the upper echelons theory, cognitive foundations rooted in demographic attributes and value systems significantly influence decision-making preferences, which, in turn, affect corporate innovation outcomes [1]. The concept of faultlines emerges as a pivotal construct for assessing top management team characteristics, reflecting the heterogeneity of groupings based on distinct attributes within the team [2]. As digitalization and intelligent technologies reshape organizational processes, innovation activities increasingly involve cross-departmental, inter-organizational, and even cross-border collaborations. Therefore, it is essential that we fully consider the impact of faultlines on innovation in the context of digital intelligence.

The existing research on the subsequent effects of top management team characteristics has produced relatively rich results, but it often focuses narrowly on a single feature. Based on previous studies of heterogeneity and diversity, Lau and Murnighan [2] introduced the concept of “faultlines” defined as a virtual dividing line that separates a team into several subgroups based on one or more attribute characteristics. Members with similar characteristics tend to cluster together, increasing the similarity within subgroups while exacerbating the divergence between them [2,3]. Earlier heterogeneity and diversity studies often adopted an individual-level perspective. Instead, faultlines consider the interactions of different attributes—they emphasize a combination-based approach [4]. These combinations, based on shared characteristics, significantly influence the realization of governance effectiveness in top management teams [5].

In the context of digital intelligence, information has become a strategic resource for manufacturing transformation and upgrading [6]. From the information-processing perspective, diverse demographic backgrounds allow groups to leverage complementary strengths. Faultline divisions facilitate extensive exchanges of knowledge, experience, and information, aiding top management teams in efficiently integrating this information [7,8]. This leads to improved decision-making efficiency and positively impacts corporate innovation [9,10,11]. Notably, task-related faultlines, based on factors such as educational background and tenure, demonstrate this effect particularly clearly, as these elements often directly reflect differences in information-processing capabilities [7,8,12,13,14]. Conversely, social identity theory posits that individual executives classify themselves into specific groups and identify with group characteristics. This process fosters internal identity and external bias [15]. Faultlines can disrupt team unity, reduce information-processing efficiency, and hinder the achievement of efficient innovation decisions [16,17]. Relationship-related faultlines, based on factors such as gender and age, exhibit this effect more pronouncedly, as these innate factors are prone to fostering factionalism, further diminishing information-processing efficiency [18,19,20]. Therefore, the impact of top management team faultlines on innovation cannot be generalized and this paper specifically investigates the differing effects of task-related versus relationship-related faultlines. Furthermore, it is notable that the chairperson plays a pivotal role in strategic decision-making, responsible for directing the organization’s strategy and advancing key decisions. The chairperson exerts considerable influence over other executives, often shaping their compliance with strategic directives. Therefore, this study investigates the unique position of the chairperson to examine whether faultlines’ effects on innovation differ based on which subgroup the chairperson belongs to.

Additionally, prior research has demonstrated that characteristics of top management teams significantly influence corporate decision-making preferences [21]. Given the limited resources available to any company at a given time, senior executives, as representatives of investment decisions, exhibit subjective intentions closely aligned with the company’s investment direction [22]. From an information-processing perspective in the digital era, these characteristics determine whether top management teams possess long-term investment decision preferences and their ability to accurately evaluate long-term investment projects, thereby affecting corporate innovation performance. Furthermore, executives’ decisions are shaped by decision-making contexts. In the digital age, information is pervasive; companies can efficiently capture competitors’ data while maintaining transparency about their own development status. Discrepancies between expected and actual performance may expose senior executives to pressure from multiple sources. This could prompt them either to increase risk tolerance and invest in high-risk activities such as innovation or to stick to conservative approaches, both of which significantly influence the information-processing dynamics under different faultline characteristics [23,24].

The manufacturing sector is marked by active innovation and offers a substantial sample size for study. In the digital era, manufacturers must capitalize on innovation opportunities. This requires top management teams to have strategic foresight and sensitivity while integrating diverse informational resources [25,26]. Therefore, this study selects Chinese A-share listed manufacturing companies as research samples. It employs R programming to calculate faultline values and uses fixed-effects models to analyze how different types of top management team faultlines impact corporate innovation performance. Additionally, it investigates the mechanisms through which long-term investment decision preferences and performance expectation gaps influence these outcomes. The study confirms that task-related faultlines, based on educational background, tenure, professional experience, and overseas experience, facilitate efficient information exchange and sharing. These faultlines enhance information-processing quality, stimulate the team’s long-term investment preferences, and lead to improved innovation performance. On the other hand, relationship-related faultlines formed by gender and age hinder information flow, exacerbate intergroup conflicts, prevent achieving long-term investment preferences, and negatively impact innovation performance. When a performance expectation gap exists—i.e., when actual performance falls short of expected performance—executives share a common goal, that is, to reverse unfavorable circumstances. Consequently, they show more focus on long-term investments and innovations, thereby enhancing the positive effects of task-related faultlines while reducing the negative impacts of relationship-related faultlines.

The original contributions of this study are as follows: First, within the digital era, we adopt an information-processing perspective to offer fresh insights into enhancing the effectiveness of top management teams and fostering innovation vitality in manufacturing enterprises. The use of fixed-effects models enables a valuable addition to the existing research by examining how top management team faultlines influence corporate innovation decision-making. Secondly, this study demonstrates the mechanisms through which top management team faultlines impact innovation performance, addressing a previously unexplored ‘black box’. This approach aids in understanding how these faultlines affect innovation decisions through long-term investment preferences from an information-processing perspective. Thirdly, by examining the pivotal role of performance expectation gaps in shaping decision contexts, we analyze, within the digital intelligence context, how senior executives as decision-makers influence innovation performance under favorable versus adverse operating conditions. These findings offer valuable insights for enhancing innovation capabilities within manufacturing enterprises.

The structure of this paper is as follows: Section 2 formulates research hypotheses based on an extensive literature review, focusing on theoretical analysis and logical explanations of variable relationships. Section 3 provides detailed descriptions of data sources, variable measurements, and the construction process of the empirical model. Section 4 presents the analysis results, including robustness tests to validate findings. Section 5 discusses the study’s conclusions and implications by integrating the existing research with practical insights from enterprises. Finally, Section 6 succinctly summarizes the core findings and outlines future research directions.

2. Theoretical Analysis and Research Hypotheses

2.1. The Main Effects of Top Management Team Faultlines on Innovation Performance

According to the upper echelons theory, the structural characteristics of top management teams significantly influence corporate performance outcomes, with faultlines serving as a critical determinant in this context [5]. Previous studies have categorized faultlines into two types based on their underlying causes: task-related faultlines and relationship-related faultlines [7]. Task-related faultlines arise from cognitive differences such as educational background, tenure, or professional experience, which can lead to disagreements regarding information processing. The diversity of demographic characteristics associated with task-related faultlines implies a broader range of knowledge, experience, and information exchange, thereby enhancing analytical capabilities and problem-solving skills while exerting a positive influence on innovation decision-making [9]. In contrast, relationship-related faultlines stem from inherent, immutable factors such as gender, age, or race, which are resistant to change. According to social identity theory, individuals tend to identify with specific groups while developing biases toward members of other groups, thereby exacerbating internal conflicts and hindering efficient innovation decision-making [7]. Therefore, this study specifically examines the differential effects of these two types of faultlines on innovation performance.

Task-related faultlines within top management teams refer to the integration of overlapping cognitive attributes among team members, examining the interactions and differences between distinct cognitive groupings [7]. A strong task-related faultline indicates both a high similarity in knowledge backgrounds among subgroup members and a clear division of the overall team based on such attributes. In the digital intelligence context, information serves as the core resource for corporate operations and strategic decision-making [27]. For top management teams, the capacity to interpret and utilize information is particularly crucial. By gathering and analyzing vast amounts of information, they can gain more precise business insights, ensuring that decisions are made with greater scientific rigor and efficiency. Research shows that individuals sharing similar backgrounds are more inclined to identify common ground, fostering enhanced communication and collaborative efficiency within these groups—this has a positive impact on decision-making processes and strategic adjustments based on information. Subgroup members view those with similar characteristics as natural allies who are more likely to support their viewpoints, creating an environment where individuals feel free to express their attitudes and ideas [8]. This is particularly advantageous for high-complexity activities like innovation, amplifying the benefits of group decision-making. From an inter-subgroup perspective, team members typically maintain an open and receptive attitude toward differing knowledge bases, demonstrating respect for the expertise of other subgroups [12]. This facilitates cross-subgroup brainstorming and idea collision based on information processing, enabling the exchange of diverse perspectives and alternative viewpoints [10,13]. Consequently, a task-related faultline fosters a ‘task-focused’ communication environment where disagreements are centered on issues rather than individuals [8,14]. For innovation-related decisions requiring careful deliberation, such faultlines provide multiple vantage points for evaluating decision feasibility while promoting knowledge and resource exchange among subgroups. This encourages the team to make creative and diverse decisions [10].

Based on the above analysis, this study proposes the following hypothesis:

H1: 

Task-related faultlines within top management teams positively influence corporate innovation performance.

In contrast to task-related faultlines, relationship-related faultlines within top management teams represent divisions based on “identity” [9]. According to social identity theory, individuals classify themselves into groups, develop preferences for their own group, and exhibit biases toward outgroup members [15]. A strong relationship-related faultline indicates both a high similarity in identity-related attributes among subgroup members and a clear distinction between subgroups based on these attributes within the team. From an intra-subgroup perspective, members tend to have a stronger psychological identification and reliance on one another, fostering greater trust and recognition within the subgroup. However, inter-subgroup differences are pronounced due to stereotyping based on demographic characteristics, leading team members to hide knowledge more frequently [18]. Consequently, knowledge and information sharing become limited to subgroup boundaries, increasing the difficulty of cross-group communication and collaboration [19]. This restricts the aggregation of collective wisdom for innovation-related decisions.

Furthermore, subgroup divisions rooted in relationship-based factors amplify boundary perceptions among members, making them prone to biases and exclusionary attitudes toward outgroup members [20]. This fosters a “people-focused” rather than an “issue-focused” dynamic within the team, where conflicts of interest arise more easily. Team members may shirk responsibilities and prioritize individual agendas over collective interests, increasing inter-group conflict within the top management team [17]. Consequently, subgroup members are more likely to adopt adversarial rather than collaborative approaches, impeding information exchange and reducing innovation efficiency.

Based on the above analysis, this study proposes the following hypothesis:

H2: 

Relationship-related faultlines within top management teams negatively influence corporate innovation performance.

2.2. Mediating Effect of Long-Term Investment Decision Preferences

Top management teams exhibit pronounced preferences for investment decisions during the decision-making process. Moderate levels of capital expenditure and R&D investments, as long-term investments, are generally recognized to contribute to corporate value realization and sustainable development. Due to their lengthy innovation cycles, high required investments, and associated uncertainties, long-term investments are often more effective in enhancing innovation performance [28]. However, motivated by self-interest, top management teams frequently prioritize short-term gains over long-term investments in practice. Therefore, exploring how to stimulate a preference for long-term investment decisions is particularly important, especially in the context of digital intelligence, where evaluating such investments is both critical and complex.

A strong task-related faultline suggests that the team possesses diverse and complementary resources, which are crucial for effective collaboration. Team members are more likely to recognize the importance of innovation and other long-term investments, adopting a forward-thinking perspective. This mindset enables them to promptly address myopic tendencies, thereby reducing opportunities for short-term arbitrage or self-serving actions. Consequently, corporate resources are redirected toward long-term investment priorities. Additionally, the team’s ability to communicate and synthesize diverse opinions internally fosters an environment where members feel supported by peers with similar characteristics [17]. This support allows them to freely and confidently express ideas, leading to enhanced decision-making through thorough discussions. These discussions facilitate detailed and accurate analyses of investment projects, enabling evaluations from a comprehensive and forward-looking perspective. Moreover, team members with varied knowledge and experience backgrounds enhance the team’s resilience against long-term risks, mitigating concerns about future uncertainties [29]. Consequently, the top management team adopts a longer-term perspective in developing corporate investment strategies, showing a clear preference for decisions that align with long-term objectives.

On the other hand, strong relationship-related faultlines create subgroup divisions based on a shared identity within teams. Members in these subgroups often exhibit significant distrust toward outgroup members [30], which undermines collaborative efforts. Consequently, investment decisions become more conservative, with top management teams favoring short-term investments that offer quick returns. This shift diverts resources away from long-term investments. Additionally, team members may avoid responsibilities, neither denying the importance of innovation nor taking ownership for strategic decisions. Instead, they may opt for risk-averse strategies, choosing short-term investments with rapid returns [31]. Such behavior leads to weaker preferences for long-term investment decisions and hinders the realization of innovation performance.

Based on the above analysis, this study proposes the following hypothesis:

H3: 

Top management team task-related faultlines enhance long-term investment decision preferences, thereby positively influencing innovation performance.

H4: 

Top management team relationship-related faultlines weaken long-term investment decision preferences, thereby negatively influencing innovation performance.

2.3. Moderating Effect of Performance Expectation Gap

The decision-making process of senior executives is frequently shaped by a multitude of factors. Given that profitability serves as a cornerstone for organizational success, the mismatch between realized performance and expected results—a concept referred to as the performance expectation gap—exerts a profound influence on both the strategic choices and operational decisions of executive teams [32]. This performance expectation gap is pivotal, as it not only determines their risk appetite but also reshapes their decision-making approaches. Within the digital intelligence context, managers establish performance expectations by analyzing a firm’s historical development and benchmarking against industry peers. They evaluate the firm’s current operational status by comparing these expectations with its actual performance. If actual performance falls short of expectations, a performance expectation gap emerges, signaling either operational challenges or a strategic misalignment within the firm. Such a decision context plays a pivotal role in shaping organizational strategies and outcomes.

From the perspective of top management team task-related faultlines, when faced with a performance expectation gap as a decision context, executives actively reflect on current challenges. They adjust their investment choices to reverse declining trends and enhance future performance [33]. A larger performance expectation gap presents greater challenges in reversing unfavorable business situations; senior executives become increasingly aware of underlying issues. They exhibit a higher tolerance for high-risk solutions and are more inclined toward long-term investments. Additionally, the performance expectation gap sends unfavorable signals to stakeholders, creating significant pressure on top management team members. Motivated by restoring their damaged reputations, executives are further incentivized to pursue long-term investments, demonstrating the company’s growth potential to stakeholders [34].

From the perspective of top management team relationship-related faultlines, the existence of a performance expectation gap signifies room for improvement in firm operations. This prompts executives to actively identify issues, analyze current conditions, and seek solutions to address poor performance. In this context, the top management team experiences a strong sense of accountability and responsibility. The shared goal of narrowing the performance expectation gap relatively weakens subgroup boundaries and reduces tensions among team members. This shift in focus directs the team’s attention toward overcoming operational challenges, resulting in a stronger long-term vision and an increased emphasis on long-term investments. Simultaneously, stakeholder oversight ensures that executives cannot overlook the necessity of pursuing such investments.

Based on the above analysis, this study proposes the following hypothesis:

H5a: 

The performance expectation gap positively moderates the positive effect of top management team task-related faultlines on long-term investment decision preferences.

H5b: 

The performance expectation gap negatively moderates the negative effect of top management team relationship-related faultlines on long-term investment decision preferences.

Given that top management team faultlines influence innovative performance through preferences for long-term investment decisions, the performance expectation gap may not only function as a moderating factor in how faultlines influence long-term investment decisions but also potentially serve as the key mechanism through which faultlines impact innovation performance via preferences for long-term investments. Therefore, this study further hypothesizes a conditional indirect effect. Specifically, the mediating role of preferences for long-term investment decisions is moderated by the decision context of a performance expectation gap. Innovation represents a crucial mechanism through which organizations address challenges. For teams with stronger task-related faultlines, when a performance expectation gap exists, members’ information communication and collaboration are more effectively facilitated under shared performance goals, thereby enhancing their pursuit of novel problem-solving approaches to a greater extent. Consequently, such teams exhibit higher preferences for long-term investments and consciously allocate resources toward innovative activities [35]. For teams with stronger relationship-related faultlines, the presence of a performance expectation gap reduces potential conflicts within the team, relatively enhancing members’ preferences for long-term investment decisions. Whether driven by self-interest or stakeholder oversight, such teams cannot overlook the importance of long-term investments as critical pathways to improving poor operational outcomes, thereby advancing innovative activities.

Based on the above analysis, this study proposes the following hypothesis:

H6a: 

The performance expectation gap positively moderates the indirect effect of preferences for long-term investment decisions in mediating the relationship between task-related faultlines and innovative performance.

H6b: 

The performance expectation gap negatively moderates the indirect effect of preferences for long-term investment decisions in mediating the relationship between relationship-related faultlines and innovative performance.

3. Research Design

3.1. Data Sources and Sample Selection

The original research sample comprises Chinese A-share manufacturing listed companies. Faultline calculations were conducted through R programming by employing hierarchical clustering combined with the pseudo-F statistic to determine the optimal number of clusters, followed by final groupings completed using the k-means algorithm. Empirical analyses were conducted via STATA (StataMP 17) software, where a Hausman test was utilized to select a fixed-effects model while controlling for the year and industry effects. The initial sample was processed as follows: (1) excluding all listed companies on the Chinese A-share market with stock codes marked as ST or *ST during 2015–2023; (2) excluding listed companies that were delisted between 2015 and 2023; and (3) removing samples with missing key variables. After these exclusions, a total of 14,888 observations were obtained. To mitigate the influence of outliers, all continuous variables were winsorized at the 1% and 99% levels.

The data used for calculating top management team faultlines and company financial metrics are primarily sourced from the CSMAR database, with missing values filled manually by referencing company websites and third-party platforms. Patent data were obtained from the CNRDS database, while R&D data related to long-term investment preferences were specifically sourced from Wind.

3.2. Variable Definition

(1)

The dependent variable is innovation performance. Patent data are commonly used to measure innovation performance. Considering the inherent lag in patent granting, this study adopts the number of patent applications as an indicator of corporate innovation performance. Given that utility and design patents exhibit relatively lower innovativeness, innovation performance is measured using the logarithm of total enterprise invention patent applications plus one. To account for the lag in patent data, the analysis employs a lagged dataset by one year.

(2)

The interpretive variable is top management team faultlines, which are analyzed following the framework proposed by Van et al. [36]. This study employs a three-step procedure to examine top management team faultlines.

Task-related faultlines are defined based on four characteristics: education, tenure in position, professional background, and international experience. Education is classified into five categories: high school or below (1), associate degree (2), bachelor’s degree (3), master’s degree (4), and doctoral degree (5). Tenure in position is calculated as the difference between the current year and the year of initial appointment, plus one. Professional background is categorized into three groups: output-oriented, production-oriented, and peripheral-oriented, each represented by dummy variables. International experience is assigned a binary value (0 or 1), where individuals with international experience include those who have studied or worked abroad. Relationship-related faultlines are defined based on two characteristics: gender and age.

Secondly, a cluster analysis is conducted based on these characteristics to group the executive team into distinct subgroups, where members within each subgroup exhibit similar features, while those across subgroups display significant differences. The clustering process involves two stages: First, the hierarchical clustering method is employed, following the methodology proposed by Ward et al. (1963) [37]; we use the sum-of-squares method and determine the optimal number of clusters (ranging from 2 to 4) for each firm through pseudo-F statistics. Subsequently, k-means clustering is applied iteratively for group allocation. A notable advantage of this two-stage clustering approach lies in its ability to partition the executive team into more than two subgroups, which better aligns with real-world business scenarios. Additionally, continuous variables are scaled based on their range rather than standard deviation, thereby preventing any single variable from disproportionately influencing the results and enhancing the overall accuracy of the analysis.

Lastly, the faultline strength and distance are calculated following the methodologies proposed by Thatcher et al. [3] and Bezrukova et al. [20]. The precise calculation formulae are detailed below:

$$Fstrengt{h}_g={\displaystyle \frac{{\displaystyle \sum _{f=1}^p{\displaystyle \sum _{k=1}^q{v}_k^g{({\overline{x}}_{kf}-{\overline{x}}_f)}^2}}}{{\displaystyle \sum _{f=1}^p{\displaystyle \sum _{k=1}^q{\displaystyle \sum _{i=1}^{v_k^g}{(x_{ifk}-{\overline{x}}_f)}^2}}}}} \mathrm{g}=1,2,\dots \mathrm{s}$$

(1)

Here, ${\overline{x}}_{kf}$ stands for the average of the k subgroup members across the f features, with ${\overline{x}}_f$ signifying the average across all members for the same features. Member i’s f feature value in the k subgroup is represented by $x_{ifk}$. Meanwhile, $v_k^g$ represents the member count in the k-th subgroup within the g grouping. The similarity within the group is measured by a value within the range of 0–1, where a higher value corresponds to a greater degree of similarity:

$$Fdistanc{e}_g=\sqrt{{\displaystyle \sum _{f=1}^p{({\overline{x}}_{f1}-{\overline{x}}_{f2})}^2}}$$

(2)

where ${\overline{x}}_{f1}$ represents the average value of the f feature for members in subgroup 1, and ${\overline{x}}_{f2}$ represents the mean value of the f feature for members in subgroup 2. The distance metric quantifies the difference between the groups, with a larger value indicating a greater difference.

Figure 1 illustrates an example where a top management team consisting of 12 members is divided into three subgroups (A, B, and C) based on characteristics 1, 2, and 3. Each small dot within the circles represents a team member, while the arrows inside indicate differences among subgroup members. The shorter the arrow, the stronger the faultline intensity; in this example, subgroup C exhibits the highest intensity, followed by subgroup B, with subgroup A showing the weakest intensity. The coarse arrows between the circles represent distances between subgroups, where longer arrows signify greater faultline distances. In this case, the distance between subgroups A and B is the largest, followed by that between B and C, while the smallest distance is observed between A and C.

Ultimately, considering the interplay between strength and distance, the faultline value is an orthogonalized interaction variable of faultline strength and distance. For clarity, we use age as a simplified practical example. Faultline strength refers to the similarity within subgroups—e.g., members aged 20, 25, and 50 forming distinct clusters. Meanwhile, faultline distance measures the relative proximity or disparity between these groups—in this case, indicating that the gap between the 20-year-olds and 50-year-olds is greater than that between the 20-year-olds and 25-year-olds. In actuality, faultlines are often more complex than this simplified example. Both strength and distance are essential components for capturing the faultlines within top management teams. To comprehensively assess the impact of faultlines, it is crucial that we consider both dimensions simultaneously rather than focusing on one at the expense of the other. By calculating the product of strength and distance, we can more effectively measure the influence of faultlines on team decision-making, ensuring neither dimension is overlooked.

Fau_g = Fstrength_g × Fdistance_g

(3)

(3)

The mediating variable is long-term investment preference (Inv-L), which is measured through long-term investment intensity. Long-term investment intensity is calculated as the combined proportion of capital expenditures (CAPEX) and research & development (R&D/RD) expenditures relative to total assets (TA). Capital expenditures include cash paid for constructing fixed assets, intangible assets, and other long-term assets, while R&D expenditures encompass both R&D expenses and capitalized R&D expenditures.

$$Inv-L={\displaystyle \frac{(CAPEX+RD)}{TA}}$$

(4)

(4)

The moderating variable is performance expectation gap (IPF), which refers to the extent to which a firm’s actual performance falls short of its expected performance. Drawing on the studies of Ref and Shapira [32], expected performance A_it is measured as a linear combination of historical expectations HA_it and social expectations SA_it. Performance is proxied by the ROA indicator, with the formula as follows:

A_it = (1 − α₁)SA_it + α₁HA_it

(5)

wherein SA_it represents the social expectations of firm i in year t, measured as the average performance of other firms within the same industry during the same year (excluding itself). HA_it represents the historical expectations of firm i in year t, measured as the average of its historical expected performance and actual performance from year t − 1. α₁ is a weight coefficient; following prior studies, we set α₁ = 0.5. Finally, subtracting the expected performance A_it from the firm’s actual performance P_it for the current year gives the difference. If the difference is less than zero, IPF is assigned the absolute value of the difference; if the difference is greater than zero, IPF is assigned a value of zero.

(5)

Control variables. To address potential confounding factors such as enterprise size, age, and financial condition, this study incorporates these variables as controls to ensure the precise estimation of the effects of explanatory variables on the dependent variable. These include enterprise size (Size), debt-to-asset ratio (Lev), cash flow ratio (Cash), operating income growth rate (Growth), enterprise age (Age), equity concentration (Top1), the year dummy variable (Year), and the industry dummy variable (Ind). The precise measurement techniques for each variable are outlined in

Table 1. Measurement methods of the main study variables.

	Variable Symbol	Variable Definition and Description
Dependent variable	Patent	The logarithm of the total number of enterprise invention patent applications is included as a variable in the model, with this variable lagged by one period to account for temporal effects
Interpretive variable	Faultline-T	The product of faultline strength and distance is calculated using educational attainment, tenure in the position, professional background, and overseas experience as predictors
	Faultline-R	The product of faultline strength and distance is calculated using gender and age as predictors
Mediating Variable	Inv-L	The combined proportion of both capital expenditure and R&D expenditure in total assets
Moderating Variable	IPF	The difference between actual performance and expected performance, and expected performance is a linear combination of historical expectation and social expectation
Control Variables	Size	The natural logarithm of end-of-year total assets
	Lev	The ratio of total liabilities to total assets
	Cash	Net cash flow from operating activities/total assets
	Growth	The growth rate of operating income, calculated as (current year’s operating income–last year’s operating income)/last year’s operating income
	Age	The natural logarithm of the number of years since establishment
	Top1	The ownership percentage of the largest shareholder, calculated as the number of shares held by the largest shareholder divided by the total number of shares
	Year	Year virtual variables
	Ind	Industry virtual variables

.

3.3. Model Construction

To test the hypotheses in this study, we specify Equation (6) for hypothesis testing. On the basis of Equation (6), Models (7) and (8) are constructed to examine the mediating effect of long-term investment decision preference, while Equations (9) and (10) serve as test models for the moderating effect of performance expectation gaps. All specifications employ a fixed-effects model:

Patent_t+1 = β₀ + β₁Faultline_t +∑Control_t + ε_t

(6)

Inv-L_t = β₀ + β₁Faultline_t +∑Control_t + ε

(7)

Patent_t+1 = β₀ + β₁Faultline_t + β₂Inv-L_t + ∑Control_t + ε_t

(8)

Inv-L_t = β₀ + β₁Faultline_t + β₂IPF_t + β₃IPF_t × Faultline_t + ∑Control_t + ε_t

(9)

Patent_t+1 = β₀ + β₁Faultline_t + β₂IPF_t + β₃Inv-L_t + β₄IPF_t × Inv-L_t + ∑Control_t + ε_t

(10)

wherein Patent_t+1 represents the innovation performance of a firm in the following year; Faultline_t denotes the top management team faultline, specifically categorized into two types: task-related faultlines (Faultline-T) and relationship-related faultlines (Faultline-R); Inv-L_t represents long-term investment decision preference; IPF_t represents performance expectation gaps; IPF_t×Faultline_t is the interaction term between performance expectation gaps and top management team faultlines; IPF_t×Inv-L_t is the interaction term between performance expectation gaps and long-term investment decision preference; Control_t represents a set of control variables affecting corporate innovation performance; and ε_t is the random error term.

4. Analysis of Empirical Results

4.1. Descriptive Statistical Analysis

Table 2. Descriptive statistics.

Variable Symbol	Mean	Standard Deviation	Least Value	Crest Value
Patent	2.6477	1.4186	0	6.5132
Faultline-T	1.3045	0.3821	0.4581	2.1356
Faultline-R	0.9521	0.2282	0.4294	1.2620
Inv-L	0.0503	0.0417	0	0.3788
IPF	0.0227	0.0356	0	0.2260
Size	22.0873	1.1590	19.9713	25.6712
Lev	0.3745	0.1827	0.0542	0.8589
Cash	0.0540	0.0643	−0.1346	0.2483
Growth	0.1612	0.3214	−0.4684	1.9284
Top1	32.7890	13.7751	8.9751	71.9153
Age	2.9586	0.2905	2.0794	3.5264

presents the results of descriptive statistics. Innovation performance has a mean of 2.6477 and a standard deviation of 1.4186, indicating relatively high levels of innovation among listed manufacturing enterprises. Task-related faultlines have a mean of 1.3045 and a standard deviation of 0.3821, while relationship-related faultlines have a mean of 0.9521 and a standard deviation of 0.2282. The descriptive statistics for the remaining variables are also reasonable.

4.2. Baseline Regression

The results of the relationship between top management team faultlines and innovation performance are presented in

Table 3. Main effect regression results.

	(1)	(2)
	Patent	Patent
Faultline-T	0.0438 **
	(2.1759)
Faultline-R		−0.0940 **
		(−2.1309)
size	0.4453 ***	0.4437 ***
	(19.0570)	(18.9919)
lev	−0.5012 ***	−0.5081 ***
	(−6.2197)	(−6.3067)
cash	0.1119	0.1272
	(0.9180)	(1.0425)
growth	0.1261 ***	0.1271 ***
	(6.1476)	(6.1920)
top1	0.0019	0.0019
	(1.2648)	(1.2887)
age	−0.3266 *	−0.3078 *
	(−1.7710)	(−1.6688)
Year/Ind	YES	YES
N	14,888	14,888
R2	0.1214	0.1214

Note: ***, **, and * indicate significance at the 1%, 5%, and 10% levels, respectively. The values in parentheses are t-values.

. Specifically, column (1) shows the regression results for task-related faultlines (Faultline-T) on innovation performance (Patent), where the regression coefficient is significantly positive at the 5% significance level. Column (2) displays the regression results for relationship-related faultlines (Faultline-R) on innovation performance (Patent), with the regression coefficient being significantly negative at the 5% significance level. These findings indicate that the presence of task-related faultlines in top management teams is significantly positively associated with innovation performance, while the presence of relationship-related faultlines is significantly negatively associated with innovation performance. This supports H1 and H2.

4.3. Robustness Tests

4.3.1. Propensity Score Matching (PSM)

To mitigate potential sample self-selection bias, this study adopts PSM to conduct robustness tests. Specifically, top management team faultlines are divided into two groups according to their median values: the higher group is designated as the treatment group, while the lower group serves as the control group. Covariates are selected from the control variables in the baseline regression model. A Logit model is employed for nearest neighbor matching with a 1:1 non-replacement ratio and a caliper of 0.05. The regression results using the matched sample data, presented in columns (1) and (2) of

Table 4. Robustness test: PSM, replacement dependent variables, and replacement sample size.

	(1)	(2)	(3)	(4)	(5)	(6)
	Patent	Patent	Patent2	Patent2	Patent	Patent
Faultline-T	0.0516 **		0.0368 **		0.0572 ***
	(2.2879)		(2.1591)		(2.6483)
Faultline-R		−0.1214 **		−0.0690 *		−0.0980 **
		(−2.3914)		(−1.8506)		(−2.0840)
size	0.4642 ***	0.4471 ***	0.4042 ***	0.4029 ***	0.4466 ***	0.4445 ***
	(17.6002)	(16.8040)	(20.4731)	(20.4109)	(17.9853)	(17.9037)
lev	−0.4516 ***	−0.3458 ***	−0.4947 ***	−0.5002 ***	−0.5168 ***	−0.5245 ***
	(−5.0317)	(−3.7449)	(−7.2657)	(−7.3481)	(−5.9867)	(−6.0764)
cash	0.1451	0.0944	0.1178	0.1293	0.1249	0.1395
	(1.0680)	(0.6715)	(1.1429)	(1.2541)	(0.9538)	(1.0640)
growth	0.1200 ***	0.1218 ***	0.1074 ***	0.1081 ***	0.1365 ***	0.1373 ***
	(5.2687)	(5.0935)	(6.1978)	(6.2358)	(6.2303)	(6.2666)
top1	0.0021	0.0024	0.0018	0.0018	0.0013	0.0013
	(1.2338)	(1.3888)	(1.4285)	(1.4532)	(0.7933)	(0.8306)
age	−0.2904	−0.2889	−0.1244	−0.1097	−0.2758	−0.2548
	(−1.4154)	(−1.3731)	(−0.7987)	(−0.7038)	(−1.4042)	(−1.2971)
Year/Ind	YES	YES	YES	YES	YES	YES
N	12,367	11,967	14,888	14,888	13,371	13,371
R2	0.1233	0.1201	0.2168	0.2168	0.1226	0.1224

Note: ***, **, and * indicate significance at the 1%, 5%, and 10% levels, respectively. The values in parentheses are t-values.

, demonstrate consistency with previous findings, indicating that the earlier conclusions exhibit high robustness.

4.3.2. Replacement of the Dependent Variable

Given that invention patents demand the highest level of technological content and innovativeness, this study adopts the number of invention patent applications as a measure of innovation performance in the baseline analysis. In robustness tests, we assign weights of 0.5 to invention patents, 0.3 to utility model patents, and 0.2 to design patents. The dependent variable is calculated as the natural logarithm of the weighted sum plus one. The main-effects regression results in columns (3) and (4) of Table 4 are consistent with previous findings, further confirming the robustness of our earlier conclusions.

4.3.3. Replacement of Sample Size

Adapting the methodology employed by Li et al. [38], this study randomly selects 90% of the original sample to form a new subsample and conducts a regression analysis using the same research framework. The results in columns (5) and (6) of Table 4 remain consistent with previous findings, further demonstrating the robustness of our earlier conclusions.

4.4. The Mediating and Moderating Effects

4.4.1. The Mediating Effect of Long-Term Investment Decision Preferences

The results testing the mediating effect of long-term investment decision preferences between top management team faultlines and innovation performance are shown in

Table 5. Test of the mediating effect.

	(1)	(2)	(3)	(4)
	Inv-L	Patent	Inv-L	Patent
Faultline-T	0.0028 ***	0.0412 **
	(3.1044)	(2.0453)
Faultline-R			−0.0038 *	−0.0904 **
			(−1.9066)	(−2.0512)
Inv-L		0.9419 ***		0.9464 ***
		(4.6690)		(4.6923)
size	0.0075 ***	0.4382 ***	0.0074 ***	0.4367 ***
	(7.1374)	(18.7312)	(7.0542)	(18.6693)
lev	0.0151 ***	−0.5154 ***	0.0147 ***	−0.5220 ***
	(4.1560)	(−6.3968)	(4.0500)	(−6.4804)
cash	0.0076	0.1048	0.0083	0.1194
	(1.3869)	(0.8599)	(1.5122)	(0.9789)
growth	0.0024 **	0.1239 ***	0.0024 ***	0.1248 ***
	(2.5637)	(6.0425)	(2.6030)	(6.0847)
top1	0.0003 ***	0.0016	0.0003 ***	0.0017
	(3.9350)	(1.0983)	(3.9728)	(1.1197)
age	−0.0061	−0.3209 *	−0.0051	−0.3030
	(−0.7292)	(−1.7416)	(−0.6115)	(−1.6441)
Year/Ind	YES	YES	YES	YES
N	14,888	14,888	14,888	14,888
R2	0.0225	0.1230	0.0221	0.1230

Note: ***, **, and * indicate significance at the 1%, 5%, and 10% levels, respectively. The values in parentheses are t-values.

. The following can be observed: In column (1), the regression coefficient of task-related faultlines (Faultline-T) on long-term investment decision preferences (Inv-L) is positive at the 5% significance level. In column (2), both task-related faultlines (Faultline-T) and long-term investment decision preferences (Inv-L) have significant positive effects on innovation performance (Patent) at the 5% and 1% significance levels, respectively. This indicates that long-term investment decision preferences play a mediating role in the relationship between task-related faultlines and innovation performance, supporting H3.

Similarly, in column (3), the regression coefficient of relationship-related faultlines (Faultline-R) on long-term investment decision preferences (Inv-L) is negative at the 10% significance level. In column (4), relationship-related faultlines (Faultline-R) and long-term investment decision preferences (Inv-L) have significant effects on innovation performance (Patent): Faultline-R has a negative effect at the 5% significance level, while Inv-L has a positive effect at the 1% significance level. This suggests that long-term investment decision preferences also serve as a mediator in the relationship between relationship-related faultlines and innovation performance, supporting H4.

Furthermore, we further employ the Bootstrap method to re-examine the mediating role of long-term investment decision preferences, conducting 500 random samplings. The results in

Table 6. Test of mediation effect by Bootstrap method.

Route		Coefficient	p	Z	95% Conf. Interval
Faultline-T–Inv-L–Patent	Indirect	0.0034	0.030	2.16	[0.0003, 0.0064]
	Direct	0.1166	0.000	5.56	[0.0755, 0.1577]
Faultline-R–Inv-L–Patent	Indirect	−0.0179	0.000	−5.54	[−0.0243, −0.0116]
	Direct	−0.3201	0.000	−10.03	[−0.3827, −0.2576]

show that the 95% confidence intervals for all paths do not include zero, confirming the existence of the mediating effect.

4.4.2. The Moderating Effect of Performance Expectation Gap

The results of the moderating effect test are shown in

Table 7. Test of the moderating effect.

	(1)	(2)	(3)	(4)
	Inv-L	Patent	Inv-L	Patent
Faultline-T	0.0028 ***	0.0394 *
	(3.0823)	(1.9546)
Faultline-R			−0.0038 *	−0.0899 **
			(−1.9285)	(−2.0408)
IPF	−0.0363 ***	−0.5968 ***	−0.0413 ***	−0.5933 ***
	(−4.1043)	(−2.7790)	(−4.6017)	(−2.7626)
Inv-L		0.9402 ***		0.9452 ***
		(4.6443)		(4.6704)
IPF*Faultline-T	0.0350 *
	(1.6603)
IPF*Faultline-R			0.0873 **
			(2.3580)
IPF*Inv-L		8.7937 *		9.1444 *
		(1.6603)		(1.7271)
size	0.0071 ***	0.4311 ***	0.0071 ***	0.4296 ***
	(6.7730)	(18.3864)	(6.7760)	(18.3251)
lev	0.0190 ***	−0.4383 ***	0.0187 ***	−0.4444 ***
	(5.0762)	(−5.2745)	(5.0089)	(−5.3497)
cash	0.0045	0.0478	0.0048	0.0624
	(0.8173)	(0.3878)	(0.8711)	(0.5058)
growth	0.0012	0.1026 ***	0.0012	0.1035 ***
	(1.2592)	(4.7916)	(1.2474)	(4.8323)
top1	0.0003 ***	0.0014	0.0003 ***	0.0014
	(3.7525)	(0.9170)	(3.8182)	(0.9358)
age	−0.0065	−0.3248 *	−0.0054	−0.3072 *
	(−0.7829)	(−1.7636)	(−0.6516)	(−1.6677)
Year/Ind	YES	YES	YES	YES
N	14,888	14,888	14,888	14,888
R2	0.0242	0.1242	0.0239	0.1242

Note: ***, **, and * indicate significance at the 1%, 5%, and 10% levels, respectively. The values in parentheses are t-values.

. In Column (1), the interaction term between performance expectation gaps and task-related faultlines (IPF*Faultline-T) is positive at the 10% significance level, with the same sign as the independent variable coefficient. This indicates that the performance expectation gap positively moderates the effect of task-related faultlines on long-term investment decision preferences. In Column (3), the interaction term between performance expectation gaps and relationship-related faultlines (IPF*Faultline-R) is positive at the 5% significance level, but with a sign opposite to that of the independent variable coefficient. This suggests that the performance expectation gap negatively moderates the effect of relationship-related faultlines on long-term investment decision preferences. These findings support H5a and H5b.

In Column (1), the coefficient of task-related faultlines (Faultline-T) is significant, and, in Column (2), the interaction term between performance expectation gaps and long-term investment decision preferences (IPF*Inv-L) is also significant. This indicates that the mediating effect of long-term investment decision preferences on the relationship between task-related faultlines and innovation performance is positively moderated by performance expectation gaps. Similarly, in Column (3), the coefficient of relationship-related faultlines (Faultline-R) is significant, and, in Column (4), the interaction term between performance expectation gaps and long-term investment decision preferences (IPF*Inv-L) is also significant. This suggests that the mediating effect of long-term investment decision preferences on the relationship between relationship-related faultlines and innovation performance is negatively moderated by performance expectation gaps. These findings support H6a and H6b.

Additionally, Figure 2 illustrates the primary research findings. The data above the path from long-term investment decision preferences to innovation performance represent the results incorporating task-related faultlines, while the data below reflect the results incorporating relationship-related faultlines. The coefficients and significance levels pointed by the performance expectation gap arrows indicate tests of the moderating effects on the relationship between faultlines and long-term investment decision preferences. Due to graphical constraints, the mediated moderation effect test has not been included here; however, it is noted that interaction term coefficients are significant at the 10% level. Detailed results can be found in the preceding discussion section.

4.5. Further Analysis Based on the Subgroup of the Chairperson

In contexts with a higher power distance, organizational members typically accept hierarchical structures and demonstrate obedience to those in higher positions. Consequently, while the chairperson’s formal role is to convene and chair board meetings, other executives frequently align with the chairperson’s influence [39]. As the top-ranking executive, the chairperson enjoys more speaking privileges and authority than other board members, enabling them to exert significant influence over innovation-related decisions.

When the chairperson is part of the majority subgroup, this subgroup tends to dominate decision-making, reflecting majority opinions while marginalizing those from minority subgroups [40]. This undermines minority subgroup members’ positivity in participating in decisions and weakens the significance of subgroup differentiation due to unequal status and imbalanced power dynamics. Conversely, when the chairperson belongs to a minority subgroup, this subgroup’s influence is significantly amplified. Despite being numerically disadvantaged, the minority subgroup gains the ability to counterbalance the majority subgroup with the support of the chairperson’s position. A balance is achieved between the traditional “minority-submits-to-majority” decision-making rule and the power dynamics influenced by the chairperson [41]. Based on this reasoning, this study hypothesizes that, when the chairperson belongs to a minority subgroup, the effects of board member faultlines on innovation performance are more pronounced.

The classification of majority and minority subgroups for the chairperson is based on the following criteria: within the 2–4 iterative subgroups identified, the subgroup with an absolute numerical majority is designated as the majority subgroup. If the chairperson is classified into this group, it is defined as a scenario where the chairperson belongs to the majority subgroup; otherwise, it is considered a minority subgroup scenario.

Table 8. Main effect test by chairperson’s grouping.

	The Chairperson Is in a Majority Subgroup		The Chairperson Is in a Minority Subgroup
	(1)	(2)	(3)	(4)
	Patent	Patent	Patent	Patent
Faultline-T	0.0212		0.0765 **
	(0.7048)		(2.1865)
Faultline-R		−0.0777		−0.1502 **
		(−1.1677)		(−2.0393)
size	0.5272 ***	0.4738 ***	0.3950 ***	0.4263 ***
	(14.2968)	(13.6862)	(10.5912)	(10.6674)
lev	−0.6660 ***	−0.6188 ***	−0.2603 **	−0.2945 **
	(−5.3505)	(−5.3116)	(−2.0029)	(−2.1563)
cash	0.1446	0.1482	0.1176	0.1960
	(0.7899)	(0.8650)	(0.6114)	(0.9874)
growth	0.1161 ***	0.1213 ***	0.1053 ***	0.1185 ***
	(3.5634)	(4.3355)	(3.3546)	(3.3723)
top1	−0.0023	0.0033	0.0077 ***	−0.0010
	(−0.9999)	(1.5047)	(3.2516)	(−0.3774)
age	−0.6939 **	−0.5497 **	−0.3128	0.0512
	(−2.4304)	(−2.0170)	(−1.0698)	(0.1610)
Year/Ind	YES	YES	YES	YES
N	7550	8126	7338	6762
R2	0.1295	0.1304	0.1247	0.1247

Note: *** and ** indicate significance at the 1% and 5% levels, respectively. The values in parentheses are t-values.

presents heterogeneity tests for task-related and relationship-related faultlines among board members according to the chairperson’s subgroup affiliation. Columns (1) and (2) display results when the chairperson is part of the majority subgroup, where neither task-related nor relationship-related faultlines exhibit statistically significant effects on innovation performance. In contrast, Columns (3) and (4) present findings for minority subgroup scenarios. Here, task-related faultlines demonstrate a positive coefficient at the 5% significance level, indicating a stronger positive impact on innovation performance. Conversely, relationship-related faultlines show a negative coefficient at the same significance level, suggesting a more pronounced negative effect. These results underscore that, when the chairperson belongs to a minority subgroup, task-related faultlines enhance innovation performance significantly, while relationship-related faultlines hinder it more substantially.

5. Discussion

This study is conducted within the context of digital intelligence, using Chinese A-Share manufacturing listed companies as the research sample. Using R 4.1.0 to measure faultlines and employing a fixed-effects model, this research examines how top management team faultlines impact innovation performance. Additionally, it explores the mediating role of long-term investment decision preferences and the moderating effect of performance expectation gaps within this framework. Given the importance of information-processing capabilities, this study enhances our theoretical understanding of how faultlines influence outcomes in a digital intelligence context. Additionally, it offers practical recommendations for manufacturing companies to build effective top management team structures and improve their innovation capacities.

In this study, the measurement of faultlines is based on the methodology proposed by Van et al. [36] and implemented using R programming. The approach involves a two-stage clustering process grounded in various attribute characteristics of the top management team. During the first stage, hierarchical clustering combined with pseudo-F statistics is used to determine the optimal number of clusters, comparing cluster numbers ranging from 2 to 4. The number corresponding to the maximum pseudo-F statistic is selected as the optimal cluster count. In the second stage, k-means clustering is applied to identify the grouping scheme based on the determined optimal cluster number. Finally, following the method proposed by Thatcher et al. [3], the interaction term between faultline strength and faultline distance is calculated to measure faultlines. This approach represents a relatively new development in existing faultline calculations. Unlike many studies that uniformly divide top management teams into two or three groups [14], this study allows cluster numbers to be flexibly determined based on corporate realities, ranging from two to four groups. This method demonstrates a greater generalizability and better alignment with practical contexts.

This study employs the Hausman test in the empirical analysis to determine the suitability of a fixed-effects model for estimating and validating the proposed hypotheses. The results confirm both H1 and H2: task-related faultlines in top management teams, formed by cognitive differences such as education, tenure, professional background, and overseas experience, have a positive impact on corporate innovation performance. Conversely, relationship-related faultlines based on innate and difficult-to-change factors, such as gender and age, negatively influence innovation performance. These findings remain robust after undergoing propensity score matching (PSM), alternative measures of the dependent variable, and sample size adjustments. Consistent with mainstream conclusions, stronger task-related faultlines under digital intelligence enhance information integration capabilities and foster more efficient collaborative interactions, thereby positively influencing innovation outcomes. In contrast, relationship-related faultlines have an adverse effect. This study provides valuable insights into optimizing top management team composition in the context of digital intelligence. To achieve efficient decision-making, enterprises should adopt techniques such as knowledge management, big data analysis, and artificial intelligence when selecting and optimizing their top management teams. They should prioritize factors representing the team’s knowledge assets, including executives’ education, tenure, professional background, and overseas experience, ensuring that the top management team exhibits a broad distribution across these factors and forms specialized subgroups based on distinct cognitive capabilities to mitigate information-asymmetry-induced decision errors. Additionally, enterprises should strive to minimize faultlines arising from identity-based divisions by consciously balancing gender and age distributions among top management members. When recruiting new executives, enterprises must carefully evaluate how potential hires may impact existing team dynamics.

Additionally, this study supports H3 and H4: the preference for long-term investments mediates the relationship between task-related and relationship-related faultlines within the top management team and corporate innovation performance. Specifically, task-related faultlines within the top management team strengthen the preference for long-term investments, which positively influences innovation performance. Conversely, relationship-related faultlines diminish this preference, resulting in a negative impact on innovation performance. This finding provides manufacturers with a basis for predicting decision-making risks associated with different types of faultlines in the context of digital intelligence. Given the limited corporate resources within a specific timeframe, long-term investments generally provide greater potential for fostering innovation. This underscores the importance of assembling top management teams with diverse qualifications (e.g., education, tenure, professional background, and overseas experience) to improve their ability to anticipate and mitigate decision-making risks. These findings suggest that manufacturing enterprises should prioritize balancing team member characteristics and emphasize executives’ long-term perspectives to maximize the beneficial effects of their preference for long-term investments.

H5a, H5b, H6a, and H6b are supported additionally. Specifically, the gap in performance expectations (as a contextual factor) influences the indirect relationship between top management team task-related faultlines, long-term investment decision preferences, and innovation performance by acting as a positive moderator. Conversely, it negatively moderates this relationship when considering top management team relationship-related faultlines. Enterprises should recognize the impact of performance expectation gaps. To some extent, adversity can act as an incentive; even significant gaps may present favorable opportunities for proactive innovation and transformative change. For organizations with substantial differences in expectations, aligning executives around a shared objective—such as overcoming operational challenges—is crucial in order to fully harness their potential. In adverse conditions, enterprises should prioritize integrating information resources and enhancing mechanisms for information sharing to foster resilience and adaptability.

Finally, considering the differences in decision-making status, this study conducted further tests based on subgroup discrepancies among chairpersons during the clustering process. Previous research often treated the influence of various executives within teams as equivalent; however, in reality, the chairperson, as the highest leader of the enterprise, can directly impact or even determine innovation-related decisions. The findings reveal that, when a chairperson is part of a minority subgroup, both task-related and relationship-related faultlines within the top management team have more significant effects on innovation performance. This suggests that chairmen should avoid unilateral decision-making and refrain from exerting authority to dominate discussions. Instead, they should ensure the active involvement of members from different subgroups in all stages of the decision-making process. In particular, under the context of digital intelligence, enhancing communication channels is crucial in order to ensure decisions are based on thorough discussions among executives. This approach not only improves the scientific nature of decision-making but also fosters a more inclusive and collaborative environment within the organization.

In terms of theoretical implications, considering the context of digital intelligence, this study contributes theoretically by examining faultlines from a combined perspective, moving beyond individual attributes. Through the lens of information integration and processing, the research investigates how task-related and relationship-related faultlines influence innovation performance. This analysis enriches the literature on top management team faultlines, particularly within the manufacturing industry, while deepening our understanding of how such teams shape innovation decisions. Methodologically, the study considers both faultline strength (reflecting within-group homogeneity) and faultline distance (capturing between-group differences), categorizing executive teams into two to four subgroups. This approach introduces a novel analytical dimension to the field. Additionally, as key decision-makers in corporate investment strategies, executives’ subjective preferences significantly influence a company’s strategic direction. Given the resource constraints inherent in any organizational context, the efficiency of information processing becomes critical. By introducing “long-term investment decision preferences” as a mediating variable, this study seeks to unpack the relationship between executive team characteristics and innovation performance. Furthermore, the research evaluates variations in the effects of executives’ preferences for long-term investment decisions and innovation performance under contrasting business conditions—specifically, whether there is a gap in performance expectations.

In practical terms, this study’s findings yield actionable recommendations for manufacturing firms. When assembling or optimizing executive teams, companies should prioritize members’ heterogeneity across diverse dimensions. By fostering effective information integration and communication, organizations can assemble teams that are better aligned to support innovation, maximizing the benefits of faultline advantages while minimizing potential drawbacks. The findings underscore the importance of prioritizing the structural composition in executive team design and improving the information-processing efficiency. Additionally, by enhancing the long-term vision, companies can adopt a broader perspective on strategic decision-making, enabling choices that better align with innovation goals. These insights offer practical relevance for organizations seeking to embed innovative practices into their operations. Moreover, the study’s analysis of performance expectation gaps provides guidance for manufacturing firms in optimizing team structures based on specific contextual factors, such as whether the organization experiences gaps between anticipated and realized performance outcomes. This approach facilitates more efficient resource allocation and innovation-related activities, ultimately supporting firms’ long-term competitiveness and sustainable growth.

6. Conclusions

This study employs a two-stage clustering approach to identify faultline groupings and measures their effects using a fixed-effects model for analysis. The research investigates the mechanisms through which different types of top management team faultlines influence innovation performance by triggering long-term investment decision preferences, as well as how these mechanisms operate under varying decision-making contexts. The findings demonstrate that task-related faultlines have a positive impact on innovation performance, while relationship-related faultlines exert a negative influence. Long-term investment decision preferences serve as the critical mechanism through which faultlines affect innovation performance. Moreover, performance expectation gaps amplify the positive effects of task-related faultlines and mitigate the negative impacts of relationship-related faultlines. Notably, when the chairperson occupies a minority subgroup position, the effects of faultlines become more pronounced.

This study makes theoretical contributions by refining the measurement of faultlines and enriching research on the downstream effects of top management team characteristics. In practical terms, it offers insights for manufacturing enterprises operating under digital intelligence, emphasizing how to fully leverage the proactive roles of top management teams. On one hand, it guides manufacturers to grasp the broader context of digital intelligence when forming or optimizing their executive teams. Specifically, they should consider differences in team members’ attributes across various dimensions, effectively utilizing the advantages of task-related faultlines in information transmission and communication while reasonably avoiding the negative impacts of relationship-related faultlines. This approach can maximize the role of top management teams in driving innovation. On the other hand, the study suggests that manufacturing enterprises should appropriately balance long-term and short-term investments, broaden their perspectives on long-term investment decision-making, and optimize team structures based on whether performance expectation gaps exist within the organization. By leveraging internal resources effectively and conducting more efficient innovative activities, enterprises can achieve better outcomes in their strategic decision-making processes.

This study has several limitations. First, although it examines top management team member characteristics such as gender, age, education, tenure, professional background, and overseas experience—commonly analyzed traits in the existing literature—it does not account for other potential factors that could influence faultline formation, such as kinship relationships and informal relationships among team members. Second, there may be omitted variables in this study, such as leadership styles and organizational culture—these factors could potentially influence the results but were not included in this analysis. Third, the innovation decision-making process of top management teams is inherently complex, suggesting there may be additional mechanisms beyond investment preferences, such as strategic choice pathways, which have not been explored in this research. Fourth, this study employs data from Chinese manufacturing firms; therefore, the generalizability of the findings to other industries or national contexts requires further discussion.

Future research directions include expanding the measurement of faultlines by incorporating more variables, such as kinship relationships, political connections, and informal relationships among team members, in order to conduct a deeper analysis. Second, the potential influence of omitted variables, such as leadership style or organizational culture, on the results can be considered. Third, a further exploration of potential mediating variables could enrich the understanding of how faultlines influence innovation performance. Fourth, future research could also explore evidence from other industries or countries to achieve a more universally applicable result.