1 Introduction

Globalization and technological revolution have compelled corporations to improve their competitive advantage, and innovation is recognized as the main engine of a firm’s long-run growth. Exploring determinants of innovation is increasingly challenging for executives and is a growing interest for economists.Footnote 1 Innovation demands knowledge, capital, technology, and human talents. These factors are not easily obtained from individuals but are much more accessible by institutions.

Theory suggests that institutional ownership is an important mechanism in influencing managers to invest efficiently and maximize firm value. For example, Leland and Pyle (1977) state that an increase in institutional ownership is a good signal on reducing asymmetric information and revealing the higher quality of the underlying project. Aghion et al. (2013) indicate that firms with higher institutional ownership are more likely to invest in innovation because financing innovative activities require a long-term and stable capital structure. However, limited empirical evidence supports this prediction. Furthermore, most of the extant researches does not clearly distinguish the types of institutional investors, and regard institutional investors as homogeneous. Whether the types of institutional investors—active and passive, influence the pace of corporate innovation differently remains an unresolved issue.

This paper is inspired by the notation that different institutional investors have different incentives to monitor their investee firms (Cremers and Nair 2005). While passive institutional investors do not exert effort in monitoring firms (Chen and Miller 2007), active institutional investors are more willing to collect private information about invested firms and provide professional advising and monitoring. As a result, if investors possess inside information about a firm, they may foster corporations to invest in high-quality, innovative projects. Consequently, institutional ownerships constitute a crucial factor that influences firm innovation.

To have inside information, institutional investors usually have a seat of directors on board. Besides monitoring, these long-term investors from institutions also provide advising function to corporation decisions including innovation activity. Beside the sophistication and voting power from institutions, the board members attributes, such as profession, gender, and ethnicity also play an important role on affecting innovation decisions. Building on the foundation of upper echelons theory, Hambrick (2007) and Berger et al. (2014) argue that experiences and personalities of executives significantly influence their interpretations of the situations they encounter and in turn, affect corporate decision. Thus, the cognitive frames of the board members are important to firm activities and outcomes. Diverse professions on boardroom influence the firm's strategic direction by providing cognitive conflict and constructive debate which may result in innovative ideas (Hillman et al. 2002; Miller and Triana 2009).

Wiersema and Bantel (1992) argue that diversity in demographic traits help bring a diversity of information sources, leading to more creative or innovative brainstorming. We thus propose that a board may affect innovation via its diversity attributes. Greater diversity in gender and ethnicity traits lead to increased monitoring, such that managerial opportunism becomes less prevalent. Besides, a growing number of banks have a seat on the board or a long-term ownership stake in the businesses of their clients also provide a prospective advising to a firm. In the US, over 30% of the largest firms have bankers on board (see, e.g., Kroszner and Strahan 2001). Ferreira and Matos (2008) explore ownership and board links of bankers worldwide and point out that institutional ownerships replace direct bank ownership gradually. Corporations exhibit relative advantages in accessing funds, advising, and monitoring when a banker has a seat on board or holds ownership. We, therefore, expect that a board with professional directors (e.g., bankers or members of the audit committees) or expertise are more competitive, assertive, and willing to engage in innovative activities.

Extant literature exploration of the relationship between ownership structure and firm innovation ignores the function of board diversity.Footnote 2 For example, David et al. (2001) examine the impact of institutional ownerships on R&D investments and suggest that institutional ownerships increase R&D inputs for short- and long-terms. By contrast, Graves and Waddock (1990) previously indicate a negative relationship between innovation activities and institutional investors, who tend to look at short-term performance. While results of ownership and innovation are mixed, the increasing demand for director diversity recently led researchers to examine diversity’s impact on boardroom behaviors. However, a salient aspect of existing literature is the absence of studies on the effects of board diversity on innovation.

This study narrows the gap in the literature by providing a novel perspective at how institutional ownerships and board diversity within the modern enterprise may help improve firm innovation in the US. US listed firms make an excellent testing ground for investigating this topic because of their high proportion of institutional ownerships compared to firms of other countries. In fact, institutions held 70% of the equity in US firms during our sample period, indicating a compound annual growth rate of 3.9% over the last 15 years. Furthermore, unlike other countries where data are trivial, unobtainable or of poor quality, the US offers access to high-quality data on institutional ownerships, board diversity, innovation investment, and patents.

A concern of this study is that institutional ownerships are not exogenous random variables. Instead, they are endogenously affected by many factors (Demsetz and Lehn 1985; Aggarwal et al. 2011). To address issues related to the endogeneity of institutional ownerships, we used the instrumental variables approach. We first employ a firm's presence in theStandard and Poor’s 500 (SP500) index as an instrument to instruct institutional ownership.Footnote 3 Duggal and Millar (1999) indicate that institutional ownership significantly determined by a firm's presence in this index. SP500 is representative for regularly listed firms and is relatively stable over time. As such, many institutional investors prefer to invest in firms included in this index. Besides, institutional investors also steadily hold shares of large (high market-value) firms because large capitalization is associated with high liquidity and investment safety. Following Elyasiani and Jia (2010), we use firm capitalization as the second instrument. SP500 and firm capitalization are good instruments given that they are likely to influence institutional ownerships but are unlikely to impact firm innovative activities directly.

Furthermore, we examine the effects of board diversity by the proportion of ethnic minorities, the percentage of the audit committee, and if the board includes female appointees. Additionally, this research focuses on firm R&D expenditure and combines that information with the NBER Patent Database, USPTO and IPTECHFootnote 4 Patent Databases. The patent count could connect the output of R&D activities and innovation investment (Deng et al. 1999). Moreover, biases due to firm heterogeneity and endogeneity problem are a concern in our analysis. We choose a set of control variables and 2SLS regression to address the endogeneity issue. The difference-in-differences (DID) analysis and an alternative measure and methodology are conducted for the robustness check. The results remain similar.

Our findings reveal a positive relationship between institutional ownership and R&D investment: firms with high institutional ownership have significantly higher R&D investment and innovation output as measured by the number of patents. In particular, we find that the positive relation between institutional ownership and innovation is mainly driven only by active institutional investors rather than by passive institutional investors. Noticeable, the impact of these passive institutional investors turn out to be significantly positive on innovation when a banker has a seat on the board. Our findings also indicate that firms with more directors serving as audit committee members, higher proportion of ethnic minority, or females on the board have higher innovation investment and patents.

Finally, earlier literature recognizes that highly innovative firms face difficulties in attracting equity capital, especially, institutional investment (Bushee 1998; Graves and Waddock 1990), due to the high level of information asymmetry. However, the Sarbanes–Oxley (SOX hereafter) Act in 2002 is strongly believed to reduce information asymmetry, thus benefiting all firms (Engel et al. 2007). Our results show that while the SOX Act benefits innovations of all firms, passive institutional investors benefit more than active ones. Thus, the SOX Act narrows the gap between active and passive investors in innovation investment and reduces the competitive advantage of active investors in innovation. In addition, this study suggests that the implementation of the SOX Act mitigates the differences between active and passive institutional investors. Further, this effect erodes the differential impact on innovation of active versus passive investors.

The remaining parts of this paper is organized as follows. In Sect. 2, we briefly review relevant literature. Section 3 describes the data and methodology, followed by empirical results in Sect. 4 and the conclusions in Sect. 5.

2 Literature review

Innovation is the process of developing new technological knowledge to generate a higher-quality or lower-cost product than those previously available (O’Sullivan 2000). Many prominent scholars have consistently stressed innovation's importance as the key to economic development and business growth (Cohen and Klepper 1996; Zahra and Covin 1995). In fact, innovation activities considerably benefit enterprises, but are often cited as risky investments with high probability of failure and uncertain return. Therefore, an innovation investment is an investment decision that may generate conflicts between shareholders and managers.

Agency theory suggests that managers are risk averse due to concerns about their undiversified human capital (Fama 1980), and literature investigates how managers can be motivated to make risky choices through various corporate governance mechanisms. These mechanisms comprise both internal and external methods, such as monitoring by the board of directors and shareholders. Existing literature has increasingly focused on the role of corporate governance mechanism in influencing firm performance (Munari et al. 2010). This study sheds new light on the literature by investigating how institutional ownership and board diversity influence firm innovation.

2.1 Institutional ownership and innovation

An emerging stream of literature is examining the causes and effects of institutional ownership and revealing a continuing expansion of their role in corporate governance. Institutional investors have become the dominant investors in the financial markets of many countries, and have become prevalent as means of a common collective ownership type. With their sophistication and high ownership, institutional investors are more likely to monitor and discipline managers (Elyasiani and Jia 2010). Thus, such investors play a vital role in influencing managers to invest efficiently and maximize firm value (Bushee 1998). An empirical study by David et al. (2001) find that large ownership stakes held by institutional investors grant them the power to influence R&D investments. Similarly, Hoskisson et al. (2002) test the relationship between governance and corporate innovation strategies and uncover evidence that professional investment fund managers preferred acquiring external innovation.

Conversely, Graves and Waddock (1990) reveal a negative relationship between innovation activities and institutional investors who would look more at short-term performance. Using a sample of Korean firms, Choi et al. (2012), determine that institutional ownership has a positive effect on firm innovation performance. In a study of 303 Chinese hi-tech manufacturing businesses, Jiang et al. (2013) report that institutional ownership has a positive relation with internal R&D activities, but a negative relationship with innovation performance (new product). A general implication of these findings is that institutional investors equipped with strong information-processing capacity and voting power can motivate top managers to pursue innovation projects with prospects. Rong et al. (2017) employ Chinese data and find that the effect of institutional investors on firm patenting mainly comes from mutual funds. Chang et al. (2019) found that higher institutional ownership leads to more innovations, including higher citations and patents. However, existing literature examining this issue views institutional investors as homogeneous. We explore this issue by separating the institutional investors into active and passive types and adding board diversity details for more governance insight.

2.2 Board diversity and innovation

The board of directors is one of the key internal corporate governance mechanisms to “control agency problems and mitigate information asymmetry between the firm and outside stakeholders” (Fama and Jensen 1985). While many researchers study the influence of board characteristics on firm performance, studies on its impact on firm innovation are limited. For example, Tseng et al. (2013) find a positive relationship between board size and innovation ability. Dong and Gou (2010) uncover evidence that the presence of independent outside directors and director ownership leads to superior innovation investments. However, these studies treate directors as a homogenous group without controlling for their personal characteristics, such as ethnicity, gender, and qualifications. Building on the upper echelons theory (Hambrick and Mason 1984; Hambrick 2007), we argue that variations in these characteristics may explain the difference in innovative effort or risk-taking investments among firms.

2.2.1 Gender diversity

Gender diversity is defined as the percentage of women in the board and proxy by the gender composition in boardroom (Marinova et al. 2015; Byoun et al. 2016). Literature has reported a drastic increase in the number of female directors and professionals in the last few decades around the world. Women now are getting the important education and performing their duties well. If the female directors' qualifications and caliber are confirmed then gender diversity is one of the major signals of board diversity (Hafsi and Turgut 2013). A study for US firms between 1993 and 1998 shows that board gender diversity leads to improved firm performance by raising return on assets (Erhardt et al. 2003). Governance reforms indicate the importance of gender diversity in boardroom (Adams and Ferreira 2009).

Gender diversity are effective because female directors tend to take their roles very seriously in boardrooms and show less problems of attendance (Singh and Vinnicombe 2004). Besides, boards with gender diversity are provided with a better understanding of the marketplace thus possible to have better decision-making capabilities (Carter et al. 2003). More notably, Adams and Funk (2012) find that female directors exhibit increased sensitivity to social and environmental issues and are more risk-seeking than their male counterparts. Deore et al. (2021) find that board gender diversity is positively associated with the quantum, impact, and risk of innovations. If one believes that the investment in innovation projects is risky, we would expect that gender diversity of the board may moderate the impact on firm innovation.

2.2.2 Ethnic diversity

While the nationality of corporate directors around the world is gaining increasing attention, race and ethnicity seem to have become important dimensions of board diversity in the US. The role of diversity in board composition is well documented, and the emphasis has been on both gender and ethnicity. Carter et al. (2003) find a significant positive relationship between the proportion of ethnic minority directors on boardroom and Tobin’s Q. Similarly, by using a sample of firms in Norway and Sweden, Oxelheim and Randøy (2003) determine a significantly higher value for enterprises that have outsider Anglo-American board members.

Tseng et al. (2013) argue that variety could bring firms more flexibility in today's volatile environment, and the diverse ethnicity of directors is benefits firms by encouraging better investment decision. Ethnic diversity may bring different viewpoint on boardroom (Hillman et al. 2002), break familiar investment patterns, which then opens up a wider range of strategic options to be considered and increased the awareness of innovation and opportunity. We therefore argue that the higher diverse ethnic minority, the better to stimulate broader discussion for alternative considered, which may facilitate innovation activities of firms. We thus expect that ethnic diversity of boards increases corporate innovation.

2.2.3 Qualification diversity

The board of directors is entrusted with crucial firm decisions, and the quality of decision-making is likely to depend as much on their qualifications, experiences, and skills. Monks and Minow (1995) find that director expertise and occupational characteristics may affect the board’s ability to monitor management and enhance firm performance. Raghunandan et al. (2001) recommend strengthening director qualifications and highlighting the crucial role of internal auditors in assisting audit committees in the internal control process. Similarly, Darmadi (2013) examines the effect on financial performance of the educational backgrounds of the directors and the CEO, and concludes that educational qualification is not always a good proxy for managerial quality. The author suggests several factors that need to be considered, such as managerial skills, experiences, networks, and other skills obtained beyond academia.

Dewally and Peck (2010) find that professional directors who are members of audit committees or have previously worked for the government, universities, or business associations, are viewed as management human capital assets of the firm. However, limited research exists on the impact of professional directors on firm innovation. If the qualifications of the director are associated with greater monitoring, then managerial opportunism becomes less prevalent. Therefore, we expect that these high-quality directors will become more competitive, assertive, and more willing to be risk-taking regarding investment in innovation projects.

3 Data and methodology

3.1 Data sources

To construct the sample for this study, we combine data from several sources. The institutional ownership data are obtained from the Thomson Reuters Institutional (13F) Holdings. SEC requires all institutional organizations, companies, universities, and so on, to exercise discretionary management of investment portfolios over US$100 million in equity assets and to report those holdings. All common stock positions greater than 10,000 shares or US$200,000 must be reported. As noted on the WRDS website, the type code variable on the Spectrum is not reliable after 1998. We then follow Bushee (1998) and Bushee et al. (2010)Footnote 5 in taking the “reliable” Spectrum type codes and we carry these data forward in time for institutions still in existence after 1998. The information on corporate boards is from the Risk-Metrics database (formerly Investor Responsibility Research Center), which covers S&P 500, S&P Midcap 400, and S&P SmallCap 600 firms for 1996 to 2014. As information on committee membership for 1996 and 1997 are missing, we exclude those years.

Two databases are utilized for information on the innovation inputs and outputs pursued by sample firms. First, innovation input data are obtained from the Compustat. We define R&D investment (RD_SALE) as the R&D expenditure divided by total sales, measured at the end of fiscal year t. Second, the innovation output data measured by the number of patents were from the Harvard Patent Network Dataverse, which has updated the NBER-2006 database to 2010 (Li et al. 2014). To augment the data, we employ other Patent Databases from Intellectual Property Technology Innovation System (IPTECH) and USPTO to update USA patent data from 2011 to 2013. The above databases provide detailed information on US patents, such as patent assignee names, the number of patents, and the grant year. We use utility patents as a proxy for innovation because it is known as “patents for invention” in the US.Footnote 6 We measured innovation output by the natural logarithm of 1 plus the number of patents granted [LN (1 + PATENTS)]. Following He and Tian (2013), we add one to the actual values when calculating the natural logarithm of the number of patents to avoid losing firm-year observations with zero patents.

Finally, we collect firm characteristics (such as firm size, age, cash ratio, leverage, return on assets, and growth opportunity data) from the Compustat Fundamentals Annual database as control variables (explanations and definitions are presented in the following section). After merging the various data sources, we impose three restrictions on the data. First, firms operating in financial sectors (SIC codes 6000–6999) are excluded because they are subject to different regulatory accounting considerations. Second, we exclude all firm-year observations with missing values for explanatory variables. Third, following previous studies, we exclude all variables at the 1st and 99th percentiles to alleviate the effect of outliers (Aivazian et al. 2005; Cleary 1999). Consequently, the final dataset includes 13,565 firm-year observations from 1998 to 2014. Table 1 provides descriptive statistics of the variables.

Table 1 Summary statistics
Full size table

3.2 Measuring the variables

This section provides the definitions of the dependent and independent variables. Detailed variable definitions are presented in Appendix A.

3.2.1 Innovation

Based on the input and output of innovation activities, we develop the following three proxies for the degree of firm innovation: (i) ratio of R&D expenditure to total sales (RD_SALE), (ii) natural logarithm of 1 plus the number of patents [LN (1 + PATENTS)] registered by a sample firm, and (iii) innovative efficiency, measured by the ratio of patents relative to R&D capitalization. First, we measured innovation input using R&D intensity made by a firm during the fiscal year. We define RD_SALE as the ratio of R&D expenditure (Compustat item XRD) to total sales (Compustat item SALE). Second, we measured innovation output by by the natural logarithm of 1 plus the number of patents granted to firms [LN (1 + PATENTS)]. Following He and Tian (2013), we add one to the actual values when calculating the natural logarithm of the number of patents to avoid losing firm-year observations with zero patents. Finally, to address relative input (research effort, such as R&D expenses)/output (patent numbers) problems, we employ innovative efficiency (IE) as another dependent variable. Innovative efficiency in year t is measured as patents granted in year t scaled by the R&D capitalization in years t-2 to t-6. Following the definitions of Hirshleifer et al. (2013), the formula of IE are specified as below.

$${IE}_{i,t}=\frac{Number\,\, of \,\,{Patent}_{i,t}}{({RD}_{i,t-2}+0.8 x {RD}_{i,t-3}+0.6 x{RD}_{i,t-4 }+0.4 x{RD}_{i,t-5 }+0.2 x{RD}_{it-6})}$$

where RDi,t-2 denotes R&D expenditure (Compustat item XRD) in fiscal year ending in year t-2, and so on.

3.2.2 Institutional Ownerships

Total institutional ownerships (IO_TOTAL): IO_TOTAL is the ratio of institutional ownerships to total shares outstanding at the end of each fiscal year. Following Ferreira and Matos (2008) and Aggarwal et al. (2011), we set institutional ownerships to zero if the firm is not held by any institution in the Thomson Reuters Institutional (13F) Holdings. According to David et al. (2001), institutional investors care about the stability of the firm and focus on long-term investment to maximize their benefits. With sophistication and significant shareholdings, institution investors have the power to influence the allocation of scarce resources for competitive and challenging investments, such as innovation, and monitor how investments are being utilized. We, therefore, expect a positive relationship between institutional ownership and firm innovation.

Institutions, which are long-term investors, could serve as board members to monitor managers for mitigating information asymmetry and influencing the decisions of innovation activity. Following Bushee (1998) and Bushee et al. (2010), we classify “the different types of investors into active and passive institutional investors. Active institutional investors are investment companies, independent investment advisors, and public pension funds. Passive institutional investors include insurance companies, private pension fund, and others”. Active institutional ownership (IO_ACTIVE) is the sum of the holdings of all active institutional investors to total shares outstanding at the end of each year. Passive institutional ownership (IO_PASSIVE) is the sum of the holdings of the ownership by passive institutions to total shares outstanding. As mentioned, while passive institutional investors do not exert effort in monitoring their investee firms (Chen and Miller 2007), active institutional investors are more advantageous because of their strong information-processing capacity and their willingness to gather private information about investee companies. Thus, if investors know inside information, they may foster corporations to invest in high-quality innovative projects.

3.2.3 Board diversity

Qualification diversity (AUDITING): we use AUDITING, which is measured as the percentage of directors serving on the audit committee to characterize professional directors in the firm. Theoretical literature suggests that higher quality management teams may invest in long-run value oriented projects. Given that innovative projects are among these long-run value-enhancing projects, we expect that a board composed of members with higher skill, experience, and expertise will invest more in innovative projects and have greater extent of innovative output.

Ethnic diversity (ETHNIC_MINORITY): we consider ethnic diversity by the proportion of the ethnic minority (African-American, Hispanic, and Asian) on the board. Prior literature suggests that diverse ethnicity of members of the council would bring different perspectives, promote better understanding of the cultural and market place, and offer new ideas for problem solving (see e.g., Carter et al. 2003). Therefore, we expect a positive relationship between ethnic diversity and innovation.

Gender diversity (FEMALE): we use a female appointment dummy variable (FEMALE), which is set to one, if the firm has at least one female director on board as a proxy for gender diversity. Women are argued as “paying more attention to communication, collaboration, personnel development, and networking” (Claes 1999). Moreover, female directors display increasing sensitivity to social and environmental issues and are more risk-seeking than their male counterparts (Adams and Funk 2012). If such notions are true, then we expect that women would be more efficient in working with senior management to directly enhance firm innovation.

3.2.4 Control variables

Our choice of control variables is motivated by their potential relevance as noted in prior literature. Our control variables fall into two different categories: board and firm characteristics.

Board characteristics: Following previous studies, we control a series of board characteristics that may influence the innovation capacity of a firm. These variables include board size, board independence, and CEO duality. Board size (BOARD_SIZE) is measured by the number of directors serving on the board. Literature indicates that the functioning of a board can affect the quality of managerial decision and firm performance (Fama and Jensen 1983). We predict that a larger size of the board of directors will result in better performance in terms of firm innovation.

Board independence (BOARD_INDEPEND) is measured by the percentage of independent directors on the board of directors of the company. Independent directors are usually from universities, research institutions, and law firms. According to the characteristics of this group, Baysinger and Hoskisson (1990) prove that “independent-outside directors could improve R&D investment in companies given their long-term orientation”. Therefore, we expect that independent directors have a positive effect on firm innovation. CEO duality (CEO_DUALITY) is defined as a dummy variable that equals 1 when the chairman of the board also serves as CEO, and 0 otherwise. A dual CEO benefits the firm if CEO works closely with the board to create firm value (Brickley et al. 1997). However, such situation makes it easier for the him or her to assert control of the board and consequently make more difficult for shareholders in terms of monitoring and disciplining the management (Lehn and Zhao 2006). Based on these arguments, we predict an ambiguous relation between CEO duality and the firm innovation investment efficiency.

Firm characteristics: To isolate the effect of institutional ownership and board diversity on innovation output, we control firm characteristics documented as important innovation determinants by previous studies. The first control variable is firm size (SIZE), which is measured by the natural logarithm of total assets (Compustat item AT). Firm size reflects the present and the prospects for innovation (Craig and Dibrell 2006). Hall and Ziedonis (2001) argue that large firms and capital-intensive firms undertake more innovation activities. We expect that firm innovation positively correlates with firm size.

Return on Assets (ROA) captures profitability. We define ROA as the net operating income (Compustat item NI) divided by the book value of total assets (Compustat item AT). Hitt et al. (1991) finds a negative relation between ROA and patent intensity. However, Fang et al. (2014) report that firm profitability has a positive effect on the level of firm innovation.

Firm Leverage (LEVERAGE) is the ratio of total debt (Compustat item DLC + DLTT) to total assets (Compustat item AT). Generally, bank managers would require collateral for innovation loans. If firms do not obtain sufficient cash inflow and need external debt financing, they may show less innovative projects (Czarnitzki and Kraft 2009). Accordingly, we expect firm leverage to be negatively related to firm innovation.

We further include sale growth (GROWTH) as proxy for firm growth opportunities. GROWTH is measured as the average of the total sales (Compustat item SALES) growth over the sample period. Increase or decrease in sale growth provides a signal of the firms’ innovation activities. Therefore, we expect a positive relationship between sales growth rate and innovation. Cash ratio (CASH RATIO) is measured by total cash (Compustat item CH) to total assets (Compustat item AT). Cash ratio shows the percentage of company assets held in cash and marketable securities. CAPEX_TA is the capital expenditure ratio, which is measured by capital expenditure (Compustat item CAPX) divided by total assets. A large capital expenditure might indicate that such firm has significant growth opportunities. We then expect that cash ratio and capital expenditure ratio are positively related to R&D investment. We employ two-digit SIC dummies to control for the industry effects.

3.3 Empirical models

This study investigates the nature of the relationship between the institutional ownerships, board diversity, and firm innovation. Such investigation entails regressing firm innovation (INNOVATION) in corporate operations on variables that capture institutional ownership, board diversity influence, and controlling for other board characteristics, firm characteristics, industry effect, and event factors. The regression specifications are as follows:

$$ INNOVATION_{it} = \beta_{0} + \beta_{1} INSTITUTION_{it} + \beta_{2} BOARD \, DIVERSITY_{it} + \beta_{3} CONTROLS_{it} + \varepsilon_{it} $$
(1)
$$ Where \;INSTITUTION_{it} = \alpha_{0} + \alpha_{2} CONTROLS_{it} + \alpha_{3} INSTRUMENTS_{it} + \omega_{it} $$
(2)

Subscripts i and t indicate firm and time, respectively. Variable definitions are as follows: INNOVATION is a proxy for firm innovation in corporate operations. INSTITUTION is a collection of proxies that capture the power of institutional investors (such as institutional ownership in total, active institutional ownership, and passive institutional ownership) to influence corporate innovation investment decisions. A concern is that institutional ownerships are not exogenous random variables, but are endogenously affected by many factors (Demsetz and Lehn 1985; Aggarwal et al., 2011). To address issues related to the endogeneity of the institutional ownerships, we use the instrumental variables approach.

We first employ the Standard and Poor’s 500 index (SP500_D) as an instrumented variable (see, e.g., Duggal and Millar 1999). Standard and Poor’s is an index that is representative for regularly listed firms and is relatively stable over time. As such, many institutional investors prefer to invest in firms included in this index. A dummy variable SP500_D is used, which is set to 1 if the firm is in the S& P 500 index and 0, otherwise. SP500_D is likely to influence institutional ownership because fund managers are typically benchmarked against this index, but is unlikely to have direct impact on firm innovation. Additionally, institutional investors prefer to steadily hold large stocks (high market-value) because large capitalization is associated with high liquidity and investment safety (Elyasiani and Ja, 2010). Hence, we use market capitalization as the second instrument. Market capitalization (LNMKVALT) is measured by the log of the firm value (Compustat item MKVALT).

We first build a model using the above instrumental variables and all exogenous variables in the innovation equation to predict institutional ownerships. Given that equations are estimated using the same data, their error terms may be correlated. Therefore, we adopt a 2SLS regression to address the endogeneity issue and correlated errors between equations. The first stage model is shown as Eq. (2). A fitted value of institutional ownerships, computed by using first-stage estimates, is used to replace the observable institutional ownerships in the second stage. The second stage model is shown in Eq. (1):

$$INNOVATION_{it} = \beta_{0} + \beta_{1} INSTITUTION_{it} + \beta_{2} BOARD \, DIVERSITY_{IT} + \beta_{{3}} CONTROLS_{it} + \varepsilon_{it} \qquad (1)$$

where, BOARD DIVERSITY is a collection of variables that capture AUDITING (percentage of directors who also are audit committee members), ETHNIC_MINORITY (proportion of the ethnic minority in the board), and FEMALE APPOINTMENT dummy variable. CONTROLS are a set of variables comprising control board and firm characteristics, as well as industry effects. We specifically control for board size, board independence, CEO duality, firm size, profitability (ROA), growth opportunities, cash ratio, capital expenditure ratio, and industry dummy variables.Footnote 7 ε and ω are the error terms.

We employ the Hausman specification test (Hausman 1978) to confirm the existence of endogeneity. Further, the identification test and excluded-instruments F- test provide the check for the soundness and adequacy of instruments. Table 2 presents the Hausman test results. We first regress institutional ownership variables on the selected instrumental variables and the rest of the exogenous variables as the model in Eq. (2). The initial regressions of institutional ownership variables against instrumental and exogenous variables resulted in a p-value for instrumental variables that are small enough to conclude that SP500_D and LNMKVALT are the best instruments. After that step, the residuals of institutional ownerships (IO_TOTAL, IO_ACTIVE and IO_PASIVE) equations are plugged, one by one, into the original regression of Eq. (1). The results in Columns (2), (4), and (6) of Table 2 show that these residuals are statistically significant. Furthermore, the Hausman test for endogenous of IO_TOTAL, IO_ACTIVE, and IO_PASIVE show the F-value of 485.94, 20.25, and 6.47 with p-value of 0.000, respectively. These results indicate that IO_TOTAL, IO_ACTIVE, and IO_PASIVE are endogenous in terms of their relationships with firm innovation, and thus 2SLS is necessary and is justified.

Table 2 Hausman test for endogenous problem
Full size table

4 Empirical results

4.1 Descriptive statistics

Table 1 Panel A provides descriptive statistics for the full sample. We present the means, medians, standard deviations, Q25, and Q75, for key variables. The average firm has a R&D investment of 10.37% to total sales and 5.07% of total assets. The median of RD_SALES is 2.66%, demonstrating that US firms have spent much on R&D investment, and R&D varies widely across firms. Table 3 Panel A illustrates the R&D distribution showing that more than 20% of firms have no R&D expenditure, and few firms have high R&D ratio. Most of R&D distributions are less than 10%. Table 3 Panel B reports industry distribution. In terms of R&D to total sales, four sectors represent more than 10.37%: agriculture, computers, durable manufacturers, and pharmaceuticals. These industries also have high ratio of R&D to total assets.

Table 3 Distribution of firm innovation, institutional ownerships, and board diversity
Full size table

The interesting aspect of institutional ownership is that on average, 70% of firm equity is held by institutional investors. Six industries that have high proportion of institutional ownership are computers, extractive, mining and construction, retail, and services. Active and passive institutional investors hold 54% and 19% of the equity, respectively. These results show that this sample firm has high institutional ownership and most firms are active investors. These results further confirm the important role of institutional investors in financing the capital for US firms, as shown in existing literature. Table 2 Panel C shows the time distribution result, which is in line with our prediction, that institutional ownership (both active and passive ones) is higher after the Sarbanes–Oxley Act in 2002. In particular, the mean of institutional ownership before 2002 is about 58% compared with 65% after 2002. The time trend of R&D investment and institutional ownerships during the sample period are illustrated in Figs. 1 and 2, respectively.

Fig. 1
figure 1

Institutional ownerships over years, 1998–2014

Full size image
Fig. 2
figure 2

R&D and institutional ownerships over years, 1998–2014

Full size image

Moreover, the results show that the board size of average firms is 9.29 with about 10% of female directors. The average firm has 14% ethnic minority (African-American, Hispanic, and Asian) in its board and 20% are in the audit committee board. The percentage of independent directors for the median firm is about 70% and 67% of CEOs serve as board chairs. The average firm size (LNTA) is 7.358, which indicates that most firms have huge amounts of assets. ROA of the median firm is 3.32%. The average cash ratio is 0.095 and growth opportunities is 0.1104 with a few firms having negative or extremely high growth rate. Other control variables are displayed in Table 1 Panel A.

Table 1 Panel B presents the Pearson correlation coefficients for board diversity, board characteristic, and firm characteristic variables. Correlations among the internal governance characteristics and firm characteristics are low. This evidence suggests that each board and firm characteristic variable is potentially a candidate for inclusion in board diversity variables as a stand-alone element, rather than merely being highly correlated with other board and firm characteristic variables. We will check this relationship more carefully by running regressions.

4.2 Two-stage least squares (2SLS) regression results

We use the 2SLS regression to examine the relationship among institutional ownerships, board diversity, and corporate innovation using R&D expenditures to sales ratio (RD_SALE) as the dependent variable. Table 4 presents three regressions with different measures of institutional ownerships. Column (1) shows that total institutional ownership (IO_TOTAL) is positively associated with R&D intensity in the sample firms. In Columns (2) and (3), we split institutional investors into active and passive institutional investors (IO_ACTIVE and IO_PASSIVE). We find that active institutional ownership fosters firm-level innovation as measured by the R&D to sales ratio, unlike passive institutional ownership.

Table 4 Institutional ownerships, board diversity, and firm innovation
Full size table

In particular, the coefficient of IO_TOTAL is statistically significant at 1% level with a coefficient of 0.325. This value represents one standard deviation (0.1994) increase in total institutional ownership, which increases the magnitude of R&D investment by 0.625% [= 0.1994*0.325/0.1037] from the average R&D investment level of 10.37%. Similarly, the coefficient for IO_ACTIVE in Column (2) is 0.437, which is statistically significant. Economically, an increase of active institutional ownership increases the R&D investment ratio by 0.7674% [= 0.1821*0.437/0.1037] from its mean value. These results are consistent with the prediction that institution investors in general and active institutional investors in particular, have the power to influence the allocation of scarce resources for innovation investment and monitor the way investments are utilized. With their strong information-processing capacity and large shareholdings, the more active institution ownership a corporate has, the higher the amount of their innovation investment. Our result is in line with the findings of Bushee (1998) and Choi et al. (2011), which confirm that institutional ownership positively fosters firm innovation input. One implication is that active institutional investors are sophisticated investors who have effective monitoring roles and motivate top managers to create a force that encourages sound innovation investment.

Column (3) Table 4 reports the 2SLS regressions of firm innovation regarding the percentage of passive institutional ownership (IO_PASSIVE). A pronounced difference exists between the effect for active and passive investors. The coefficient of IO_PASSIVE is -0.111 and statistically insignificant, suggesting that passive institutional ownership is unimportant in explaining firm innovation investment. Therefore, we conclude that the positive relation between institutional ownership and innovation is mainly driven by active institutional investors.

Moreover, we find that the coefficients of our board diversity proxies (AUDITING, ETHNIC_MINORITY, and FEMALE) are significantly positive, regardless of which institutional ownership variable is included. This finding also implies that better audit quality, higher proportion ethnic minority directors, and a female appointment to the board all result in improving board monitoring and advising, with added willingness for risk-taking investment and positive effect on innovation investment. Specifically, the magnitudes of these effects in Column (1) are as follows: 0.185 indicates that one standard deviation (0.0748) increase in audit quality leads to a 0.1334% [= 0.0748*0.185/ 0.1037] increase in R&D investment; 0.232 represents that an increase in standard deviation (0.0888) of the proportion of ethnic minority directors would increase innovation investment by 0.1986% [= 0.0888*0.232/ 0.1037]. Similarily, the coefficient of FEMALE is 0.028, indicating that a female appointment to the board increases R&D investment by 0.27% (= 0.028/ 0.1037) for business in the sample.

These results confirm the role of internal auditors in assisting audit committees in the corporate governance process (Raghunandan et al., 2001). We also control for a comprehensive set of board and firm characteristics probably affecting firm innovation. Whichever institutional ownership variable we use, control variables show signs as expected on the innovation investment, except for CASHRATIO. Evidently, larger firms, high growth opportunity firms, firms with more operating profits, and companies with a higher number of independent directors on the board are more innovative. These findings are consistent with those in existing literature (Czarnitzki & Kraft, 2009; Fang et al., 2014; Tseng et al., 2013). The overall F-statistic value in all models has a p-value of less than 0.001, indicating that the models have statistically significant explanatory power.

4.3 Robustness checks

4.3.1 The role of banker on the board

In this section, we consider an alternative measure of qualification diversity: banker appointment (BANKING_COMMIT) on the board. We use BANKING_COMMIT as a dummy variable, which equals to 1 if a firm has at least one banker on board providing professional banking services, or 0 otherwise. As noted by Kroszner and Strahan (2001), over 30% of the largest US firms have bankers on their boards. Given the potential link between a banker on the board and solid innovation investment, we examine the benefits of bank monitoring in firm management for innovation by adding banker appointment dummy into Eq. (1). We also add an interaction between INSTITUTION and BANKING_COMMIT to capture the incremental effect of firm innovation investment response of the banker appointment to its institutional ownership. We estimate the following equation:

$$ \begin{aligned} INNOVATION_{it} = & \beta_{0} + \beta_{1} INSTITUTION_{it} + \beta_{2} INSTITUTION_{it} \times BANKING\_COMMIT_{it} \\ & + \beta_{3} BANKING\_COMMIT_{it} + \beta_{4} BOARD \, DIVERSITY_{it} + \beta_{5} CONTROLS_{it} + \varepsilon_{it} . \\ \end{aligned} $$
(3)

Table 5 reports the results of the 2SLS in Eq. (3). The results indicate a positive relationship between active institutional ownerships and R&D investment, whereas such relationship is negative for the passive ones, thereby confirming our earlier results. The key coefficient corresponding to interaction terms (β2) are positive, suggesting that the effects of institutional ownerships are more pronounced when firms have a banker on the board. Table 6 reports the estimation summary to compare the effect of active and passive institutional ownerships between firms with a banker and firms with a non-banker on the board. The result highlights that active institutional ownerships have tangible positive effects on firm innovation, even when firms have no banker on the board. For passive institutional ownership, the result is a negative effect on innovation investment but a positive influence once firms have a banker on the board.

Table 5 Institutional ownership and banker on the board
Full size table
Table 6 Estimation summary of banker on the board
Full size table

The coefficients for IO_PASIVE and IO_PASIVE × BANKING_COMMIT are -0.301 and 0.418, respectively, and both are significant at the 1% level. Regarding economic magnitude, one standard deviation (0.0731) increase in passive institutional ownership reduces the magnitude of R&D investment by 0.212% [= 0.0731*-0.301/ 0.1037] from the average R&D investment level of 10.37%. However, one standard deviation (0.0731) increase in passive institutional ownership increases R&D investment by 0.082% [= 0.0731*(0.418—0.301)/ 0.1037] for firms with a banker on the board. This evidence indicates that a banker on the board can change the impact of passive institutional ownership on innovation investment from negative to positive.

Our findings align with Byrd and Mizruchi's (2005) idea, indicating that when bankers serve as providers of professional services, their expertise regarding capital markets can be precious. With a banker on the board, firms could easily borrow funds from the banks with lower spreads or better non-price loan terms (Francis et al. 2012). Moreover, facing a profitable investment, bankers could foster innovation investment through equity finance for these innovative projects.

4.3.2 Alternative measures of innovation

We also explore whether our findings are robust to an alternative measure of innovation. We consider two other alternative measures of firm innovation: innovation output and innovative efficiency for this check. Innovation output was measured by the natural logarithm of 1 plus the number of granted patents [LN (1 + PATENTS)], whereas the innovative efficiency (IE) in year t measures the patents granted in year t scaled by the R&D capitalization over the years t-2 to t-6. Results in Table 7 show that institutional ownership- and board diversity-related variables remain robust except AUDITING. The significance and signs are similar compared to the prior results. In particular, the IO_TOTAL coefficients are 2.28 and 1.169 on innovation output and innovative efficiency, respectively. The magnitudes of these effects are economically significant, indicating that one unit increase in the standard deviation of institutional ownership (0.1994) leads to a rise of 12.01% in the mean of the number of granted patents and 0.206% [= 0.1994*1.12169/ 1.1305] in the average of innovative efficiency (1.1305). Similarly, the coefficient on IO_ACTIVE is 1.15 for innovative efficiency, suggesting that one standard deviation increase in active institutional ownership (0.1821) creates a 0.203% [= 0.1821*1.150/ 1.1305] increase in innovative efficiency from its mean value. Furthermore, the coefficients of IO_PASSIVE are -0.835 in the innovation output equation and − 0.690 in the innovative efficiency equation; both are statistically significant. This finding verifies that passive institutional ownership has a negative impact on both innovation output and a firm’s innovative efficiency. Our analysis indicates that institutional investor is an important determinant of corporate innovation, and active institutional investors mainly drive this relation.

Table 7 Alternative measures of innovation
Full size table

Evidently, board diversity positively correlates with firm innovation, confirming our earlier results. Besides, nearly all the significance and signs remain except the AUDITING where a negative effect on innovation output appears. Similarly, while bank provides resources and expertise to firms to foster innovation, it motivates innovation expenditure rather than innovation outputs in the same period because patent outputs take a long time to invent. Table 5 Column (3) shows that the coefficient of the interaction between the bank and passive institutional ownership on R&D is positive and significant, with a magnitude of 0.418; however, it is negative on patents. Furthermore, we note that the coefficients of other control variables are highly stable.

4.3.3 The effect of the Sarbanes–Oxley Act 2002

Previous studies indicate that the highly innovative firms face difficulties in attracting equity capital, especially institutional investment (Bushee 1998; Graves and Waddock 1990). One of the main reasons is the high level of information asymmetry between managers and investors, which translates into high monitoring costs for outside shareholders (Jensen and Meckling 1976). The Sarbanes–Oxley Act was strongly believed to generate improvements in terms of financial disclosure and quality of information (Engel et al. 2007). Thus, we examine whether the enactment of SOX Act in 2002 has an effect on the relationship between the institutional ownership and firm innovation investment. Accordingly, we employ the multivariate difference-in-differences (DID) analysis with the 5-year window centered on the event year.

An SOX2002 is a dummy variable, which equals 1 if the year is after 2002; and 0 otherwise. We must first identify a control group unaffected by SOX2002. Despite the mandatory implementation of the SOX Act, the Securities and Exchange Commission allowed small firms (market values of equity under US$75 million) to have more time to comply with the Act (Zhang et al. 2007). On the contrary, the treatment group is an accelerated filer, which includes companies having the market value of equity higher than US$75 million. We create a dummy variable (ACCELERATED_FILER), which equals to 1 if a firm is an accelerated filer; and 0 otherwise. We add the interaction terms to Eq. (1), which estimates the following equation:

$$ \begin{aligned} INNOVATION_{it} = & \beta_{0} + \beta_{1} INSTITUTION_{it} + \beta_{2} INSTITUTION_{it} \times SOX2002 \\ & + \beta_{3} INSTITUTION_{it} \times ACCELERATED\_FILER + \beta_{4} INSTITUTION_{it\;t} \times SOX2002 \\ & \times ACCELERATED\_FILER \, + \beta_{5} SOX2002 + \beta_{6} ACCELERATED\_FILER \\ & + \beta_{7} BOARD \, DIVERSITY_{it} + \beta_{8} CONTROLS_{it} + \varepsilon_{it.} \\ \end{aligned} $$
(4)

We first add the interaction of INSTITUTION × SOX2002 to capture the differential effects of institutional ownership on R&D investment due to the enactment of SOX Act. We also add the interaction of INSTITUTION × ACCELERATED_FILER to compare the effect between the accelerated filer and non-accelerated filer groups. Finally, we interact SOX2002, ACCELERATED_FILER, with INSTITUTION, resulting in the following interacted terms: INSTITUTION × SOX2002 × ACCELERATED. We expect that the relationship between innovation investment and institutional ownership variables (IO_TOTAL, IO_ACTIVE, and IO_PASSIVE) are more pronounced for the treatment group over the post-SOX period. The “clean” effect of institutional ownership on innovation arises from the exogenous change in institutional ownership forced on the accelerated filer through a mandate to disclose accurate and complete information.

Table 8 reports the DID estimation results with the 5-year window centered on the event 2002, which confirms again that total institutional ownership positively relates to firm R&D investment. The coefficient is significant with an even higher magnitude compared with that in Table 4. Interestingly, although the interaction of SOX2002 and passive institutional ownership over the post-SOX period turns out to be significantly positive, the interaction of SOX2002 and active institutional ownership is negative. These findings imply that although the SOX Act requires listed firms to disclose accurate and complete information, the advantages from active institutional investors are reduced. Instead, the passive investors gain significant benefits from the reduction in firm information asymmetry and have increased willingness to invest in highly innovative firms.

Table 8 Evidence from a Quasi-Experiment: The SOX Act Impact
Full size table

In addition, the positive and significant coefficient of IO_TOTAL × SOX2002 × ACCELERATED_FILER illustrates the marginal effect of SOX2002 on the institutional ownership–innovation investment relation for accelerated filers. We also find similar results with active and passive institutional investors on their relation with firm innovation for accelerated filers. The results support the perspective that the enactment of Sarbanes–Oxley Act mitigates the level of information asymmetry between businesses and investors (Zhang et al. 2007). Given a reduced degree of information asymmetry post-SOX, highly innovative firms can attract more institutional investments in the post-SOX period. Especially, we find higher effects of passive institutional investment in innovative firms in post-SOX.

An underlying assumption of the DID analysis is that the two groups of firms in comparison have similar characteristics or that the two groups of firms had a parallel path before the SOX2002 event. We find that this condition is met. We conducted t-tests to examine whether the innovation investment and firm characteristics in the year immediately before the event are different across the accelerated and non-accelerated filer. The results in Panel B of Table 8 reveal no significant differences in innovation investment and firm characteristics in terms of board diversity, sale growth, cash ratio, firm leverage, and capital expenditure ratio between the two groups. These results provide more confidence that the DID estimate reflects a causal effect of SOX200 on corporate innovation investment.

Table 9 reports the DID estimation summary for a time comparison before and after the SOX enactment between the non-accelerated filer and accelerated filer. For the non-accelerated filer groups, the results support our expectation that the sign effect of institutional ownership (IO_TOTAL, IO_ACTIVE, and IO_PASSIVE) remains after the enactment of SOX. For the accelerated filer group, the results show a significant change. In particular, one standard deviation (0.1821) increase in active institutional ownership increases the magnitude of R&D investment by 0.607% [= 0.1821*0.346/ 0.1037] from the average R&D investment level of 10.37% per-SOX and increase by 0.899% [= 0.1821*0.512/ 0.1037] post-SOX. By contrast, one standard deviation (0.0731) increase in passive institutional ownership reduces the magnitude of R&D investment by 0.432% [= 0.0731*(− 0.604)/0.1037] from the average R&D investment level of 10.37% per-SOX. However, the sign changes post-SOX such that one standard deviation (0.0731) increase in passive institutional ownership increases the magnitude of R&D investment by 0.052% [= 0.0731*(0.074)/0.1037] from its mean value. The results suggest that the SOX Act benefits all the investors but benefits more from passive institutional ownership. Thus, the SOX Act narrows the gap such that previously large competitive advantage in terms of information asymmetry is narrowed down. The study suggests that as the SOX Act is increasingly implemented, the differences between the active and passive institutional investors diminish; this reduction erodes the differential influence on innovation of active versus passive investors. In sum, the effect of institutional ownership on innovation is robust over the DID model specifications and controls.

Table 9 DID estimation summary
Full size table

4.3.4 Complementary analysis of gender diversity

In this section, we consider an alternative measure of gender diversity: woman director ratio (GENDER). We measure women director ratio as the number of female directors on a firm's board scaled by board size. Results are presented in Table 10 Column (1). The coefficient estimates on GENDER is positive and significant at the 10% level, suggesting that gender diversity is positively associated with corporate innovation investment. This finding is consistent with the idea of Low et al. (2015), indicating that more gender diverse boards add value by improving board monitoring and bringing new perspectives to the board, which are beneficial to firm investments. Other key estimates are again consistent with the results in Table 4.

Table 10 Alternative measures of board diversity and quantile two-stage least squares
Full size table

We further examine the role of active and passive institutional investors in innovation where women are present on the corporate boards. In this respect, we add the interactions between INSTITUTION and FEMALE to create Model (5). The remaining independent and control variables are the same as in the previous model:

$$ \begin{aligned} INNOVATION_{it} = & \beta_{0} + \beta_{1} INSTITUTION_{it} + \beta_{2} INSTITUTION_{it} \times FEMALE_{it} \\ & + \beta_{3} BOARD \, DIVERSITY_{it } + \beta_{{4}} CONTROLS_{it} + \varepsilon_{it} \\ \end{aligned} $$
(5)

Passive institutional investors are less likely to serve as a board committee (Chen and Miller 2007) and they welcome even small improvements in governance relative to firms. Prior research mentions the value of board diversity, suggesting that female directors offer diverse viewpoints, promote lively discussion to the boardroom (Letendre 2004) and transparency (Upadhyay and Zeng 2014). Therefore, we would expect a gender diversity foster more innovation investment in firms while institutional investors are passive. On the contrary, active institutional investors are more likely to have a seat on board to monitor managers and influence the corporate decision (David et al. 2001). It could be that active investors would bring flexibility and new perspectives on investment into the boards to maximize their interests, leading to higher innovation investment. We expect that the benefits of control by active institutional investors would be higher in firms with gender diversity.

Table 11 summarizes the results of complementary analyses. We analyze the interacted effect of the gender diversity and institutional ownership on innovation. The results remain similar on the impacts of IO_TOTAL, IO_ACTIVE and IO_PASSIVE on RD_SALE, including both signs and significances. The IO_ACTIVE × FEMALE presents a negative effect on RD_SALE (with a coefficient of − 1.499), statistically significant at 1% level, while IO_PASSIVE × FEMALE has a positive effect at the same significance (with a coefficient of 2.504). It reconfirms that active institutional investors in firms without female gender show a stronger impact on corporate innovation investment (with a coefficient of IO_ACTIVE = 1.563) than those with female gender on the board (coef. IO_ACTIVE + coef. IO_ACTIVE × FEMALE = 1.563–1.499 = 0.064).

Table 11 Complementary analysis of gender diversity
Full size table

As for the passive institutional ownership, the impact on firm innovation is negative, it is particularly negative in firms without female gender on board (coef. IO_PASSIVE = coef. IO_PASSIVE + coef. IO_PASSIVE × FEMALE = − 2.540 + 2.504 = − 0.036). Therefore, the positive impact of the active institutional investor on innovation is wide in firms without gender diversity on board. On the contrary, the negative impact of passive investors on innovation is significantly narrow for firms with a female on the board. Our results confirm the presence of diverse directors on the board in fostering corporate innovation. Nonetheless, the advantage of gender diversity to innovation is more significant for passive institutional investors than active investors.

4.3.5 An alternative methodology

We again perform the Instrumental Variable Quantile Regression (hereafter IV-QR) to address the concerns of the patent distribution problems for additional robustness check. Its key difference between IV-QR and OLS and 2SLS lies in the following: instead of solving for a “conditional mean,” the quantile model solves for a “conditional quantile”. The IV-QR is the same as the original quantile regression, except for incorporating instrumental variables to account for endogeneity (Chernozhukov and Hansen 2008). Unfortunately, no official Stata commands exist for this model, compelling us to follow Asongu and Kodila-Tedika (2015) and utilize the naïve quantile regression in this paper. The coefficient estimate results are divided based on quantiles and presented in Table 10. The results indicate an increasing coefficient of institutional ownership as the quantile increases. The results also show a significant different effect of institutional ownership between low and high quantile.

Our results of the relations between institutional investors, board diversity, and firm innovation are robust after implementing the difference-in-differences analysis, IV-quantile regression, and alternative measures of firm innovation.

5 Conclusions

This paper examines how institutional investors and board diversity jointly reshape the role of corporate governance and influence a firm’s innovation. Using a large sample of US firms, this research obtains some interesting results below.

First, we find institutional ownerships positively influence the innovation only appears in the active institutional investors, not in the passive investors. However, the passive investors positively affect innovation once those firms have at least one banker on the board. The plausible explanation is that a banker on the board benefits firms from professional banking advising, monitoring, and funds support with lower spreads or better non-price loan terms (Francis et al. 2012). Moreover, bankers could foster investment by lending large loans to finance innovative projects.

Second, enacting the Sarbanes–Oxley Act in 2002 facilitates the provision of accurate and transparent information to investors, narrowing the gap between active and passive institutional investors in innovation investment. Our results from the difference-in-differences analysis show that both types of institutional investors benefit from innovation post-SOX Act. However, passive investors gain more benefits from the SOX Act than active investors and become positively affected innovation investment after the event.

Third, a board with a female, a higher presence of auditing committee, or a higher proportion of ethnic minority directors significantly lifts the R&D investment and produces more patents. These findings imply that board diversity is beneficial in fostering innovation. Again, firms with a high percentage of independent directors significantly boost more investment in R& D.

Our findings support the view that institutional investors are not simply attracted to firms with high growth rates but also play a critical role in promoting corporate growth through innovation investment. These institutions are active in enhancing firms’ governance mechanisms and innovation effectiveness.