Do Long-Term Institutional Shareholders Always Vote in Favour of Board Recommendations? The Moderating Effect of Cash Holdings

Abstract

This article aims to examine the voting behaviour of long-term institutional shareholders towards board recommendations on management proposals and resolutions and how the potential agency costs could moderate such voting behaviour. This study is conducted using all corporate capital proposals put to vote by management during the annual general meetings (AGM) of publicly listed firms on the London Stock Exchange over a period of 17 years from 2000 to 2016. Building on agency theory and the concept of the monitoring function of institutional shareholders, this study finds that long-term institutional shareholders do support board recommendations on management proposals, but potential agency concerns linked to excess cash holding can negatively moderate this relationship. Additional analysis reveals that this moderating effect is observed only for management proposals related to cash inflows, specifically after the 2007–2009 financial crisis. This study highlights the importance of long-term institutional shareholders actively monitoring firms’ cash holdings and using voting to address agency concerns while advising corporate managers to optimise cash management and stay attuned to shareholder preferences. For policymakers, the research suggests promoting transparency in corporate governance and strengthening shareholder engagement to reduce agency problems and improve governance. Several robustness tests are conducted, and the results support our predictions.

1. Introduction

Prior research on corporate governance has documented the importance of institutional shareholders’ engagement in maximising firm value, and shareholder voting was found to be an effective channel used by institutional shareholders for that purpose (Iliev et al. 2015). Indeed, shareholder voting is not only considered to be one of the most powerful channels that shareholders can use to engage in influencing investee firms’ decisions (Mallin and Melis 2012; Yermack 2010) but is also considered to be an expressive tool for shareholders’ evaluation of firm performance (Tsukioka 2020) and board’s elections and composition (Michaely et al. 2023), even if the shareholders’ vote is not aligned with the majority vote (Sauerwald et al. 2013).

Recent evidence, however, shows that institutional shareholders have a heterogeneous voting behaviour (Brav et al. 2024; Brickley et al. 1988; Gine et al. 2017; Gordon and Pound 1993). The cost-benefit trade-offs in collecting firm-specific information are considered an important indicator of shareholders’ willingness to monitor managers and evaluate firm performance (Chung et al. 2018). In theory, shareholders with benefits exceeding costs would be encouraged to be effective monitors. Previous literature suggests that shareholders with a long-term investing horizon gain monitoring benefits that exceed the costs (Bushee 1998; Chen et al. 2007) and are more concerned about long-run firm value than the momentum gain from increases in share price (Bushee 2001; Derrien et al. 2013; Harford et al. 2018; Luo and Xu 2024). Supporting this argument, a survey study by McCahery et al. (2016) shows that long-term shareholders intensively and actively engage with firm management through different channels in regard to corporate governance matters.

On the one hand, as long-term institutional ownership is considered an effective corporate governance mechanism (Ayres and Cramton 1993; Harford et al. 2018), average shareholders will have more confidence in the board’s recommendations as long-term institutional ownership increases. Moreover, since long-term shareholders maintain their monitoring function for a long period, management should better understand shareholders’ expectations and consider such expectations when making their recommendations (Stathopoulos and Voulgaris 2016a). Thus, this study aims to answer the question of whether the ownership of long-term shareholders influences the voting outcome on the board recommendations on management proposals. It is generally expected that long-term institutional ownership increases positive voting on board recommendations.

In line with agency theory, long-term institutional shareholders are expected to change their voting behaviour when there is an increase in agency concerns. McCahery et al. (2016) document that while long-term institutional shareholders intensively intervene in firm decisions, they are mostly triggered by agency issues (e.g., “corporate fraud”, “inadequate corporate governance” and “low payments to shareholders despite high cash holdings”). This study aims to examine if agency concerns affect the voting of long-term shareholders on the board’s recommendations on the proposals put to vote by management. To test the dynamic voting behaviour of long-term institutional shareholders with the agency issue, the cash deviation from the optimal level is used as an agency trigger that can cause a change in shareholder voting. Holding a high level of cash increases agency concerns (Jensen 1986), which can motivate long-term institutional shareholders to put pressure on management to reduce the firm’s excess cash holdings (Cleary and Wang 2017; Huang and Petkevich 2016). Thus, as a firm holding excess cash can raise agency concerns, we expect that long-term institutional shareholders would vote negatively on the board’s recommendations as a means to express their dissatisfaction.

The research problem examines how institutional shareholders’ voting behaviour is influenced by their investment horizon and agency concerns, specifically focusing on firms’ excess cash holdings. It investigates whether long-term shareholders are more likely to vote in favour of management proposals under normal circumstances but shift to dissent when agency issues, such as excessive cash, signal inefficiencies in corporate governance. The findings of the present paper support the expectations regarding the dynamic voting behaviour of long-term institutional shareholders. The study finds that, on average, long-term shareholders vote in favour of the board’s recommendations. Yet, this relationship is moderated by the level of cash that firm managers choose to hold. The findings show that long-term institutional shareholders tend to vote against the board’s recommendations when the firm has excess cash.

This study is expected to contribute significantly. First, it finds that institutional shareholders have heterogeneous behaviour towards voting based on their investment horizon. While Stathopoulos and Voulgaris (2016a) study the effect of investor horizon on the voting outcome, our study is distinct in three aspects. This study focuses on all corporate capital resolutions put to the vote by firms’ boards rather than only focusing on Say-on-Pay resolutions. In the UK, shareholder voting on Say-on-Pay became mandatory in 2002 and was made binding in October 2013 (Stathopoulos and Voulgaris 2016b). Thus, this study aims to enrich our understanding of the voting behaviour of shareholders with a long-term investment horizon regardless of whether the voting on resolution is regulated or not. This study also differs from Stathopoulos and Voulgaris (2016a) by using the actual investment duration of a shareholder in a particular firm as a proxy for the investment horizon rather than the churn ratio of a shareholder’s overall portfolio. Building on recent literature (Attig et al. 2012; Elyasiani and Jia 2010; Wang 2014), this study adopts investment holding duration as a measure, as it more accurately reflects the investment horizon at the firm level rather than across an investor’s entire portfolio.

Second, our study also extends our understanding of the dynamic nature of institutional shareholders’ voting behaviour given the different contexts of firms’ financial policies. We argue and find that the voting outcomes of long-term shareholders vary when firms hold more cash than the optimal level. This finding seems to provide evidence for the importance of cash policies for institutional shareholders. While prior studies provide evidence that firms adjust their cash holdings toward optimal levels (e.g., Orlova and Rao 2018), our study shows that institutional investors use voting on management proposals as a channel to improve cash policies and decrease excess cash.

Lastly, this paper provides new evidence on the changes in the voting behaviour of institutional shareholders. Previous literature documents an increasing trend of activism and dissent voting after the recent financial crisis of 2007–2009 (e.g., see Stathopoulos and Voulgaris 2016a for UK firms; Sato and Takeda 2023 for Japanese firms). While our initial results show that institutional shareholders increase (decrease) their dissent (positive) voting on management proposals when a firm incurs higher potential agency costs, holding excess cash, further analysis reveals that this effect only exists after the financial crisis. These results provide evidence that illustrates that institutional shareholders consider firms’ corporate policies and use their voting as a discipline/influencing tool to incorporate corporate changes.

The rest of the paper is structured as follows. The next section presents a review of related literature and the development of hypotheses. Then, the data sources, variable construction, and econometric models used are discussed in Section 3. After that, Section 4 presents the results and provides some robustness checks. Section 5 discusses the findings of the paper. Finally, the conclusion of the paper summarises the main findings, provides recommendations and implications, and discusses several limitations and potential future research venues.

2. Literature Review and Hypotheses Development

2.1. The Importance of Shareholder Voting

During the last three decades, many pieces of legislation have been introduced to improve firms’ accountability and transparency and to mitigate corporate agency concerns (Conyon and Sadler 2010). Recent reforms not only encourage shareholders to vote during an AGM but also give them binding voting power to shape important corporate decisions (Iliev et al. 2015). For instance, legislation that makes the voting outcome on directors’ remuneration binding has been adopted in different countries to empower shareholder voting (Iliev et al. 2015). Shareholders around the world effectively use their right to vote as a governance mechanism to voice their opinions about various corporate issues (Iliev et al. 2015). Indeed, the voting channel is a fundamental aspect of corporate governance (Becht et al. 2016) and one of the most powerful instruments that enables shareholders to engage with the boards of their investee firms (Mallin and Melis 2012). Further, Sharfman (2020) argues that shareholder voting could be considered “the most prominently debated corporate governance issue of recent times”.

Yermack (2010) reviews recent research on shareholder voting and shows that firms’ boards pay considerable attention to shareholder voting and conduct significant changes in corporate governance and strategy based on the voting outcome. Alissa (2015) shows that boards not only consider shareholders’ voting decisions but also respond “selectively” and “swiftly” to shareholder dissatisfaction. Studying shareholder voting in 43 countries, Iliev et al. (2015) show that firms’ boards directly respond to shareholder dissent by implementing more withdrawals of merger and acquisition (M&A) proposals and higher director turnover. Moreover, even the stock market reacts to shareholder dissent. For instance, Tokbolat et al. (2019) find that firms with previously high levels of shareholder dissent on M&A resolutions would have a lower announcement return on subsequent M&A deals. Indeed, shareholders not only use their right to vote as a direct mechanism to influence a firm’s decisions but also as an expressive instrument to signal their perspectives regarding particular corporate resolutions (Sauerwald et al. 2016). Voting is perceived by shareholders as an effective activism channel to enhance firm value (Brochet et al. 2021).

Institutional shareholders use their voting rights to influence a broad range of important firm decisions (Iliev et al. 2015; Yermack 2010). For instance, the previous literature documents the impact of shareholder voting on directors’ remuneration (Conyon and Sadler 2010; Gregory-Smith et al. 2014) and nomination (Iliev et al. 2015), M&A (Iliev et al. 2015; Tokbolat et al. 2019), capitalisation (Iliev et al. 2015), and anti-takeover provisions (Gine et al. 2017). In principle, shareholders engage in voting to express their standpoint regarding the status of a firm’s corporate governance and strategy (Sauerwald et al. 2016; Yermack 2010).

2.2. The Agency Theory

Agency theory provides a foundational framework for understanding the dynamics between shareholders and corporate managers. Agency theory has been extensively researched and applied across various domains of corporate governance and finance. The theory suggests that a fundamental conflict of interest exists between principals (shareholders) and agents (managers), particularly when agents have the opportunity to act in their own interests rather than in the best interest of the principals. Jensen and Meckling’s (1976) seminal work on agency costs is foundational to the theory, positing that these costs arise when the interests of managers diverge from those of shareholders. Agency theory has since been used to examine various managerial behaviours, such as risk aversion, investment decisions, and the allocation of firm resources (Eisenhardt 1989). It has been widely adopted in corporate governance research, especially in understanding the relationship between ownership structure and firm performance.

Further research extended agency theory to explore how different types of shareholders, particularly institutional investors, engage with firms to address agency problems. Studies by Shleifer and Vishny (1986) and Jensen (1986) highlight that institutional shareholders, due to their large ownership stakes, are in a position to exert influence over corporate management to reduce agency costs. Institutional shareholders are considered to have lower monitoring costs compared to individual investors, making them more effective at monitoring and engaging with firms’ management. In their survey study, McCahery et al. (2016) argue that long-term institutional shareholders are particularly motivated to monitor and influence management decisions, especially when they perceive agency problems such as inefficient use of cash holdings.

A growing body of literature examines the role of agency theory in understanding shareholder voting behaviour in response to corporate governance issues. Recent studies, such as those by Sauerwald et al. (2016), have shown that institutional shareholders, especially those with a long-term investment horizon, use shareholder voting as a mechanism to influence corporate policies that affect firm value. Agency theory helps explain why institutional shareholders are more likely to dissent when they perceive managerial decisions, such as excess cash holdings or poor governance practices, to be detrimental to long-term value. For example, the research by Dittmar and Mahrt-Smith (2007) highlights that when firms hold excess cash without distributing it to shareholders, institutional shareholders view this as a potential agency cost, leading them to increase their engagement through voting to push for better cash management policies. Furthermore, the literature suggests that shareholder activism, particularly through voting, has increased in the post-financial crisis era, reflecting a shift toward more active monitoring of managerial behaviour, as evidenced by studies such as those by Brav et al. (2024) and Tokbolat et al. (2021). These studies contribute to a growing understanding of how agency theory informs not only managerial behaviour but also the actions of shareholders seeking to minimise agency costs and improve firm governance.

2.3. Voting Behaviour of Long-Term Institutional Shareholders

In principle, shareholders would need to monitor a firm and collect firm-specific information to effectively engage and positively influence the firm’s decisions (Chen et al. 2007; Iliev and Lowry 2015; Stathopoulos and Voulgaris 2016a). Unfortunately, the effective monitoring and collecting of firm-specific information come with costs that shareholders must bear. Thus, in theory, a shareholder will only effectively engage in a firm’s decisions if the benefits obtained are larger than the associated cost of such engagement (Iliev and Lowry 2015). As long-term shareholders have “lower monitoring cost functions” and higher monitoring benefits than other shareholders (Chen et al. 2007), one would expect that long-term shareholders actively engage with firms. In line with this prediction, recent literature shows that long-term institutional shareholders actively engage with firms to influence important corporate decisions (McCahery et al. 2016; Stathopoulos and Voulgaris 2016a). Iliev and Lowry (2015) argue that shareholders with a long-term holding period are more likely to engage in voting as they bear less cost in collecting firm-specific information. Stathopoulos and Voulgaris (2016a) show that long-term institutional shareholders benefit from their ability to monitor and collect information about the firm when making their voting decisions. Recent research shows that firms with higher long-term shareholders are more likely to offer online voting to facilitate shareholder engagement (Cai et al. 2024).

As institutional shareholders have heterogeneous voting behaviour (Brav et al. 2024; Davis and Kim 2007; Gine et al. 2017; Morgan et al. 2011; Stathopoulos and Voulgaris 2016a; Tsukioka 2020), prior literature finds some evidence to suggest that firms with higher long-term institutional ownership have less shareholder dissent in their voting results. This relationship can be explained by two arguments. The first argument is related to the quality of the firm’s corporate governance. Sauerwald et al. (2016) find that good corporate governance quality at both the country and firm levels significantly decreases dissent voting. Indeed, the presence of long-term shareholders is considered to be a valuable governance mechanism (Ayres and Cramton 1993; Harford et al. 2018). Thus, considering long-term institutional ownership as an indicator of corporate governance quality, an average investor will be more likely to cast positive votes as long-term ownership increases in the firm. Sauerwald et al. (2013) find that shareholder dissent decreases with the existence of relational blockholders, and the main criterion for being a relational blockholder is to be a long-term shareholder. In line with this argument, the decreasing trend in dissent voting (Conyon and Sadler 2010) could be explained by the increase in long-term institutional ownership during recent years (Derrien et al. 2013; Harford et al. 2018).

Another explanation for the low level of dissent voting could be the continuous monitoring and engagement of long-term shareholders (McCahery et al. 2016). In a firm with more long-term shareholders, the board’s proposals will more likely be in line with shareholders’ expectations (Stathopoulos and Voulgaris 2016a). Gine et al. (2017), for instance, find that a firm’s managers are more likely to act based on the voting by pension funds compared to other types of shareholders as the pension funds have a longer holding period than other shareholders (Dittmar and Mahrt-Smith 2007). It is also logical that a firm’s management may reach a better understanding of long-term shareholders’ perceptions on how the firm is to be run, as such shareholders remain investing in the firm for a long period of time (Dressler and Mugerman 2023). Such an argument is in line with the findings of Stathopoulos and Voulgaris (2016a) which find that firms with more long-term horizon shareholders are likely to have larger positive voting results on remuneration proposals compared to firms with more short-term shareholders. Thus, considering both arguments, we hypothesise:

H1: 

Long-term institutional ownership has a significant positive effect on voting for board proposals.

2.4. Agency Concerns and Voting of Long-Term Institutional Shareholders

Iliev et al. (2015) show that shareholders increase their level of dissent voting as they experience greater agency concerns, whether these concerns are at the country level or firm-level. Sauerwald et al. (2016) argue that shareholders use dissent voting to express their agency concerns, especially in liberal market economies such as the UK. Shareholders utilise their voting opposition as a mechanism to influence the firm’s policies, such as payout policy (Lin et al. 2023) and acquisition and divestment decisions (Tokbolat et al. 2021).

Stathopoulos and Voulgaris (2016a) show that long-term shareholders are more likely to cast positive voting on Say-on-Pay if the firm does not pay excessive compensation. Indeed, shareholders with a long-term holding period are more likely to cast positive votes to support value-increasing proposals (Morgan et al. 2011). Luo and Xu (2024) show that index investors (i.e., long-term shareholders) are more likely to recall shares to vote, especially for proposals that could affect the long-term value of the firms. Agency theory is considered to be the most used theory in explaining shareholders’ attitudes towards firms’ governance issues (Hillman et al. 2011). McCahery et al. (2016) survey institutional shareholders and document that long-term institutional shareholders actively engage to influence a firm’s decisions, and they are mostly triggered by governance issues. One of the main agency triggers that is found in McCahery et al. (2016) is holding a high level of cash holdings with a low dividend distribution.

Building on the agency theory, Jensen (1986) suggests that the increase in cash holdings causes an increase in the agency problem. Recent literature empirically supports Jensen’s argument that holding cash in excess of the optimal level is associated with higher agency costs (e.g., Yao and Hong 2023); thus, adjusting cash holdings toward the optimal level lessens potential agency costs (Orlova and Rao 2018). Indeed, shareholders value corporate cash holdings at a discount in the context of weak corporate governance at both the firm level (Dittmar and Mahrt-Smith 2007) and country level (Dittmar et al. 2003).

Recent studies find that long-term institutional shareholders tend to influence a decrease in cash holdings to minimise the agency problem (Cleary and Wang 2017; Huang and Petkevich 2016). As long-term institutional shareholders effectively monitor corporate cash holdings (Cleary and Wang 2017; Huang and Petkevich 2016) and actively engage with firms to influence cash policies (McCahery et al. 2016), long-term institutional shareholders will use dissent voting as an expressive instrument (Sauerwald et al. 2016) when firms hold high levels of cash holdings. Thus, we hypothesise;

H2: 

Cash holdings moderate the relationship between long-term institutional ownership and voting for board proposals.

3. Data and Methodology

3.1. Data Sources and Sample

This study uses publicly traded firms listed on the London Stock Exchange (LSE) for the period from 2000 to 2016. The dataset is extracted from four different sources. First, shareholder voting data is obtained from Manifest Information Services Ltd. Second, ownership data is obtained from Thomson Reuters—Eikon (now Refinitiv). CEO and board data is obtained from BoardEx. Finally, DataStream is used to collect firms’ accounting data. To avoid sample bias issues, the sample period starts in 2000 as ownership data were limited before 2000 in Eikon, and it spans until 2016 to avoid the change/reaction towards investments and cash holdings policies of UK firms following the Brexit vote in June 2016, which might affect our findings. As the focus of the study is on the moderating effect of cash holdings, firms operating in industries with certain capital requirements and regulatory supervision are excluded. Thus, following the cash literature (Dittmar and Mahrt-Smith 2007; Opler et al. 1999; Ward et al. 2018), firms in financial industries (SIC code 6000–6999) and utility industries (SIC code 4900–4999) are excluded from the sample. To minimise the possibility that our results are driven by large outliers, all continuous variables are truncated at the 1st and 99th percentiles.

3.2. Model

To test the hypotheses of the study, Pooled OLS and logistic regressions with year and industry-fixed effects are employed. Following the shareholder voting literature (e.g., Gine et al. 2017; Thomas and Cotter 2007), Pooled OLS regressions with industry and year fixed effects are utilised to estimate Equation (1) for the first dependent variable, For voting, which is a continuous variable.

$$\begin{array}{ll}{Forvoting}_{it}& ={\alpha }_0+{\alpha }_1{LTIO}_{it}\ast {ExcessCash}_{it}+{\alpha }_2{LTIO}_{it}+{\alpha }_3{Excesscash}_{it}+{\alpha }_4{Firm}_{jit}\\ & +{\alpha }_5{Ownership}_{jit}+{\alpha }_6{CEO}_{jit}+{\alpha }_7{Board}_{jit}+{\alpha }_8{Resolution}_{jit}+{\alpha }_{9}time+{\alpha }_{10}industry\\ & +{\epsilon }_{it}\end{array}$$

(1)

The second main dependent voting outcome variable is a binary (Dissent), and a logistic regression is used to run Equation (2) (Gregory-Smith et al. 2014; Stathopoulos and Voulgaris 2016a).

$$\begin{gathered} {Dissent}_{it}={\alpha }_0+{\alpha }_1{LTIO}_{it}\ast {ExcessCash}_{it}+{\alpha }_2{LTIO}_{it}+{\alpha }_3{Excesscash}_{it}+{\alpha }_4{Firm}_{jit}+{\alpha }_5{Ownership}_{jit} \\ +{\alpha }_6{CEO}_{jit}+{\alpha }_7{Board}_{jit}+{\alpha }_8{Resolution}_{jit}+{\alpha }_{9}time+{\alpha }_{10}industry+{\epsilon }_{it} \end{gathered}$$

(2)

Detailed definitions and the construction of all variables are discussed in the next section.

3.3. Variables

3.3.1. Shareholder Voting

As this study aims to explore the voting behaviour of long-term institutional investors, we use the two voting outcomes that are the most widely used in the literature on shareholder voting as dependent variables. Our first dependent variable is For voting by shareholders, or in other words, voting in favour of the resolutions suggested by a firm’s board of directors. Following Iliev et al. (2015), we use the percentage of votes cast in favour of a board’s proposals over the total votes cast. The second dependent variable is Dissent voting. For dissent voting, we follow the majority of shareholder voting literature and use a binary variable that takes 1 if dissent voting reaches a particular threshold rather than adopting a continuous variable as a proxy for dissent voting (Alkalbani et al. 2019; Gregory-Smith et al. 2014; Stathopoulos and Voulgaris 2016a). The main reason for using a dummy for dissent voting is that we are only interested in the likelihood that shareholder dissent is high enough to prove that a particular management proposal is not satisfactory for shareholders when compared to other proposals. Following the literature (e.g., Alkalbani et al. 2019; Gregory-Smith et al. 2014), we use a threshold of 10% in the construction of our dissent dummy variable. So, our dissent dependent variable takes a value of 1 if the percentage sum of Against and Abstain votes cast over the total votes cast is larger than 10% and a value of 0 otherwise.

3.3.2. Investment Horizon

Our main independent variable is the investment horizon of institutional shareholders. Following recent investment horizon literature (e.g., Attig et al. 2012; Elyasiani and Jia 2010; Wang 2014; Yin et al. 2018), we use the length of time in which shareholders remain investing in a firm as a proxy for investment horizon. Given the information cost that shareholders ought to bear to engage in active voting, we only focus in this study on the voting behaviour of long-term shareholders as they have more incentives and gain higher benefits from such engagement (Chen et al. 2007). Thus, following Wang (2014) and McCahery et al. (2016), we classify an institutional shareholder as a long-term shareholder when the investment duration in a specific firm is at least eight quarters (two years). Our long-term institutional ownership (LTIO) variable is measured as the sum of the shareholding percentage of all long-term institutional shareholders in a specific period and firm.

3.3.3. Excess Cash

As we are interested in understanding how the voting behaviour of long-term institutional shareholders is moderated by firm-specific agency concerns, we use excess cash holding from the optimal cash level as a proxy of potential agency costs (Jensen 1986; Opler et al. 1999; Yao and Hong 2023). We define excess cash as the level of cash that exceeds a firm’s optimal cash level (cash_i^*). The optimal cash level could be defined as the function of several firm-level characteristics that influence a firm’s cash policies (Opler et al. 1999; Schauten et al. 2013; Huang and Mazouz 2018; Banjade and Diltz 2022). Equation (3) is used to predict the optimal level of cash, and the positive residuals of the model would be excess cash (where the firm’s actual cash holdings exceed the predicted optimal level).

$${Cash}_{it}^{\ast }={\alpha }_0+{\alpha }_1{X}_{it}+{\alpha }_{2}time+{\alpha }_{3}industry+{\epsilon }_{it}$$

(3)

where X_it is a vector of the main firm-level variables that determine the level of a firm’s cash holding (Bates et al. 2009; Opler et al. 1999), particularly firm’s size, free cash flow, industry cash flow volatility, market-to-book ratio, net working capital, capital expenditure, R&D and acquisitions. The variables time and industry are year and industry fixed effects, respectively.

3.3.4. Control Variables

In all our models, we control for four levels of data. First, we control for firm-level characteristics. Following Stathopoulos and Voulgaris (2016a), we control for firm size, leverage, market-to-book ratio, industry-adjusted return on assets (ROA) and share return over 3 years. Second, we control for main ownership-level variables. Following Conyon and Sadler (2010) and Tokbolat et al. (2019), we control for blockholder ownership. We also control for foreign institutional ownership as it is found to have a considerable impact on voting outcomes (Sauerwald et al. 2016). Third, we control for CEO and board attributes. Namely, similar to most literature (e.g., Sauerwald et al. 2016; Stathopoulos and Voulgaris 2016a; Tokbolat et al. 2019), we control for CEO tenure, CEO duality, board size, board size squared, and independent board ratio. The fourth level of control is related to resolution-level characteristics. Based on the argument that one size might not fit all for board size by Coles et al. (2008) and following Sauerwald et al. (2016), we include the square term of board size to control for the possible no-linear relationship of board size. Though board members would be a great resource for a company, the coordination and reaching a consensus would be harder for excessive board size. We use a dummy that takes 1 if a resolution has been put to vote in an emergency general meeting (EGM) rather than the main AGM and 0 otherwise (Conyon and Sadler 2010). We also control for the type of resolution as we study the voting behaviour towards a range of corporate capital resolutions (Sauerwald et al. 2016). Lastly, we control for time and industry-fixed effects in all models.

4. Empirical Results

4.1. Descriptive Statistics

Table 1. Descriptive Statistics.

	Obs.	Mean	SD	Min	P25	Median	P75	Max
	Panel A: Voting variables
For voting	11,046	0.976	0.066	0.010	0.985	0.997	0.999	1.000
Dissent	11,046	0.048	0.213	0.000	0.000	0.000	0.000	1.000
Dissent Ratio	11,046	0.018	0.055	0.000	0.000	0.002	0.010	0.965
	Panel B: Ownership variables
LTIO	11,046	0.437	0.197	0.000	0.290	0.448	0.587	0.924
Blockholding (5%)	11,046	0.159	0.114	0.000	0.068	0.143	0.228	0.667
Foreign Inst. Ownership	11,046	0.135	0.112	0.000	0.045	0.105	0.200	0.643
	Panel C: Cash variables
Cash	11,046	0.101	0.109	0.000	0.030	0.066	0.132	0.818
Excess Cash	11,046	0.426	0.494	0.000	0.000	0.000	1.000	1.000
	Panel D: Control variables
Adj_ROA	11,046	0.010	0.153	−2.290	−0.017	0.018	0.072	0.510
Share Return	11,046	0.471	1.203	−0.998	−0.289	0.232	0.905	17.704
Size	11,046	13.553	1.741	7.383	12.341	13.477	14.723	18.831
Market-to-Book	11,046	1.745	1.077	0.476	1.123	1.455	1.989	13.425
Leverage	11,046	0.198	0.163	0.000	0.056	0.183	0.297	1.114
INED in Board	11,046	0.517	0.151	0.000	0.429	0.538	0.625	0.889
Board Size	11,046	2.032	0.276	1.099	1.792	2.079	2.197	2.708
CEO Tenure	11,046	1.208	1.042	−2.303	0.642	1.386	1.932	3.091
CEO Duality	11,046	0.140	0.347	0.000	0.000	0.000	0.000	1.000
EGM	11,046	0.043	0.203	0.000	0.000	0.000	0.000	1.000

This table provides summary statistics. Panel A shows the descriptive statistics of voting outcomes variables; Panel B shows the descriptive statistics for the ownership variables; Panel C shows the statistics of cash and Excess cash; and Panel D shows control variables and cash determinates. All variables are truncated at the 1% and the 99% levels. Detailed definitions of variables are in Appendix A, Table A1.

presents descriptive statistics for all the variables we used in our model. Detailed definitions of these variables are in

Table A1. Definitions of variables.

Variable	Definition
For Voting	The ratio of for (in favour) votes to the total votes cast.
Dissent	A dummy variable equals 1 if a proposal received dissent (aginst and abstain) votes above 10% of the total votes cast, and zero otherwise.
Dissent Ratio	The ratio of against and abstain votes to the total of votes vast.
LTIO	The ratio of total outstanding shares held by all long-term institutional shareholders, with investment duration of at least eight quarters (two years).
Blockholding	The ratio of total outstanding shares held by blockholding shareholderss with at least 5% shareholding.
Foreign Inst. Ownership	The ratio of total outstanding shares held by all foreign institutional shareholders who do not domicled in the UK.
Cash	The ratio of cash and cash equivalents to the book value of assets.
Excess Cash	A dummy variable equals 1 if firm in aprticaulr year hold cash above the prediced optimal level of cash, and zero otherwise.
Adj_ROA	The ratio of net income to the book value of total assets (ROA) adjusted to industry median.
Share Return	The total stock return for the last three years prior to the general meeting date.
Size	The natural log of the book value of total assets.
Market-to-Book	The ratio of (the book value of assets—the book value of equity + the market value of equity) to the book value of total assets.
Leverage	The ratio of short and long debt to the book value of total assets.
INED in Board	The ratio of independet directors to the total number of directors on the board.
Board Size	The natural log of the number of directors on the board.
CEO Tenure	The natural log of the number of years of the CEO in the position.
CEO Duality	A dummy variable if the CEO is also the chairman of the board, and zero otherwise.
EGM	A dummy variable equals 1 if a shareholder proposal has been submitted to an emergency general meeting, and zero otherwise

. Panel A shows the statistics relating to shareholder voting outcomes. Overall, in terms of company resolutions, 97.6% of the votes cast were in favour of the board recommendations of UK firms, which is consistent with prior evidence that shows positive voting is high in the UK relative to other countries (Iliev et al. 2015). Dissent voting, on the other hand, is considerably low, with an average of 1.8% dissent voting on proposals and only 4.8% of proposals receiving dissent votes that exceed 10% of the total votes cast. Our voting results are similar to prior studies that focused on UK firms (e.g., Conyon and Sadler 2010; Stathopoulos and Voulgaris 2016a; Tokbolat et al. 2019), though these earlier studies only focused on specific types of resolution.

Panel B shows that 43.7% of UK firms’ shareholders are long-term institutional shareholders with at least 2 years of holding duration. An average UK firm is 16% controlled by large shareholders with at least a 5% holding of outstanding shares. Foreign institutional ownership represents 13.5% of an average UK firm listed on the LSE. Panel C shows that in the UK, 10.1% of corporate assets were in the form of cash holdings during the study period. While 42.6% of UK firms hold excess cash, correspondingly, 57.4% of the firms hold less cash than their optimal level. Panel D presents the summary statistics of all the control variables included in our models. Independent directors represent around 51.7% of total board members, and on average, 14% of UK firms are led by a CEO with a Chairman position. While the average industry-adjusted ROA is 1%, the three-year share return is 47%.

Table 2. Pearson Correlation Matrix.

Variables	(1)	(2)	(3)	(4)	(5)	(6)	(7)	(8)	(9)	(10)	(11)	(12)	(13)	(14)	(15)	(16)	(17)	(18)
(1) For voting	1.00
(2) Dissent	−0.65 a	1.00
(3) Dissent Ratio	−0.81 a	0.80 a	1.00
(4) LTIO	0.09 a	−0.02 a	−0.03 a	1.00
(5) Cash	−0.01	0.00	0.00	−0.18 a	1.00
(6) Excess Cash	−0.04 a	0.01	0.00	−0.19 a	0.58 a	1.00
(7) Blockholding	0.02 b	0.01	−0.01	0.31 a	0.05 a	0.00	1.00
(8) Foreign Inst. Ownership	−0.01	0.06 a	0.07 a	0.54 a	−0.13 a	−0.13 a	0.01	1.00
(9) Adj_ROA	0.06 a	−0.03 a	−0.04 a	0.22 a	−0.13 a	−0.15 a	−0.03 a	0.16 a	1.00
(10) Share Return	0.04 a	−0.04 a	−0.05 a	0.04 a	0.06 a	0.02 b	−0.04 a	0.02 b	0.23 a	1.00
(11) Size	0.05 a	0.01	0.03 a	0.44 a	−0.35 a	−0.30 a	−0.25 a	0.57 a	0.28 a	0.04 a	1.00
(12) Market-to-Book	0.02 b	−0.02 b	−0.04 a	0.03 a	0.25 a	0.18 a	−0.06 a	0.13 a	0.16 a	0.30 a	−0.11 a	1.00
(13) Leverage	0.00	0.01	0.01	0.16 a	−0.32 a	−0.20 a	−0.10 a	0.18 a	0.00	−0.08 a	0.35 a	−0.10 a	1.00
(14) INED in Board	0.03 a	0.02 b	0.03 a	0.33 a	−0.09 a	−0.08 a	−0.02 b	0.35 a	0.10 a	−0.03 a	0.40 a	0.00	0.10 a	1.00
(15) Board Size	0.03 a	−0.01	0.00	0.23 a	−0.21 a	−0.19 a	−0.21 a	0.38 a	0.21 a	0.03 a	0.65 a	0.05 a	0.21 a	0.17 a	1.00
(16) CEO Tenure	0.05 a	0.00	−0.03 a	0.05 a	−0.02 b	−0.02 b	0.05 a	−0.02 b	0.17 a	0.10 a	−0.02 b	0.09 a	−0.03 a	−0.02 c	0.03 a	1.00
(17) CEO Duality	0.02 c	−0.03 a	−0.04 a	−0.17 a	−0.04 a	0.00	−0.08 a	−0.09 a	0.02 b	0.04 a	−0.04 a	−0.02 b	−0.03 a	−0.21 a	0.03 a	−0.03 a	1.00
(18) EGM	−0.01	−0.01	0.01	−0.08 a	0.04 a	0.03 a	−0.01	−0.07 a	−0.15 a	−0.02 c	−0.08 a	−0.01	0.03 a	−0.09 a	−0.06 a	−0.03 a	0.02 c	1.00

^a p < 0.01, ^b p < 0.05, ^c p < 0.1.

presents Person correlations for all the variables employed in the main analyses. The correlation is positive between long-term institutional ownership and voting and it is negative, with dissent voting at 1%. These corrections provide preliminary support for hypothesis H1. Despite the fact that all the correlations between independent variables are below 0.6, we estimate the variance inflation factors (VIFs) to check the possibility of multicollinearity. All the individual VIFs of independent variables and the mean VIFs of all models used in our analysis are below the conventional level of 5.

4.2. Multivariable Results

As we are interested in exploring the voting behaviour of long-term institutional shareholders, we present the results of the two most used voting outcome variables. These variables are For and Dissent voting, and the results are presented in

Table 3. The voting of long-term institutional shareholders when the firm holds excess cash.

Dependent Variable =	For Voting		Dissent
	(1)	(2)	(3)	(4)
LTIO	0.020 ***	0.033 ***	−1.518 ***	−2.102 ***
	[0.007]	[0.008]	[0.538]	[0.597]
Excess Cash		0.017 **		−0.445
		[0.006]		[0.304]
LTIO X Excess Cash		−0.033 ***		1.446 **
		[0.012]		[0.621]
Blockholding	0.014	0.013	0.653	0.663
	[0.010]	[0.010]	[0.608]	[0.606]
Foreign Inst. Ownership	−0.050 ***	−0.048 ***	3.704 ***	3.536 ***
	[0.014]	[0.014]	[0.872]	[0.867]
Adj_ROA	0.005	0.005	−0.040	−0.002
	[0.012]	[0.012]	[0.385]	[0.384]
Share Return	0.000	0.000	−0.126 *	−0.132 *
	[0.001]	[0.001]	[0.073]	[0.074]
Size	0.004 **	0.004 **	−0.046	−0.046
	[0.002]	[0.002]	[0.078]	[0.077]
Market-to-Book	0.001	0.001	−0.084	−0.076
	[0.002]	[0.002]	[0.077]	[0.076]
Leverage	−0.011	−0.010	0.463	0.317
	[0.010]	[0.010]	[0.552]	[0.561]
INED in Board	0.008	0.009	−0.075	−0.083
	[0.008]	[0.008]	[0.519]	[0.519]
Board Size	0.060	0.068 *	−2.400	−2.581
	[0.040]	[0.040]	[2.315]	[2.325]
Board Size (squared)	−0.016	−0.018 *	0.543	0.584
	[0.010]	[0.010]	[0.604]	[0.605]
CEO Tenure	0.003 **	0.003 **	0.032	0.032
	[0.001]	[0.001]	[0.078]	[0.079]
CEO Duality	0.005 *	0.006 *	−0.493 *	−0.490 *
	[0.003]	[0.003]	[0.258]	[0.259]
EGM	0.008	0.008	−0.602	−0.608
	[0.006]	[0.006]	[0.394]	[0.389]
Constant	0.842 ***	0.829 ***	2.348	2.883
	[0.044]	[0.045]	[2.595]	[2.627]
Topic of proposal	Yes	Yes	Yes	Yes
Industry FE	Yes	Yes	Yes	Yes
Year FE	Yes	Yes	Yes	Yes
Adj_R2 (Pseudo_R2)	0.08	0.08	(0.18)	(0.18)
Obs.	11,046	11,046	11,046	11,046

This table presents the results of Pooled OLS and logistic regressions examining the effect of the ownership of long-term institutional shareholders (LTIO) on positive and dissent votes on proposals put to vote and the moderating effect of excess cash. Column (1) shows the relationship between LTIO and for (in favoure) voting. Column (2) Further examine the moderating effect of excess cash on the relationship between LTIO and For voting. Column (3) and Column (4) replicate the first two columns but with using Dissent voting as a dependent variable. Detailed definitions of variables are reported in Table A1. Industry and year-fixed effects are used in all models. Standard errors are robust for heteroskedasticity (i.e., non-constant variance) and clustered by firm level to deal with protentional correlation within the firm. *, ** and *** indicate statistical significance of t-test at 10%, 5% and 1% level, respectively.

.

Column (1) in Table 3 shows that For voting in corporate capital, resolutions increase with an increase in long-term institutional ownership. This means that long-term institutional shareholders tend to vote in favour of management recommendations. The finding is in line with our prediction in our first hypothesis H1. However, Column (2) suggests that the above-mentioned relationship changes with a firm’s cash deviation from the optimal level. Column (2) shows that long-term institutional shareholders reduce their For votes when a firm holds excess cash. The result of the moderating impact of cash deviation on the voting behaviour of long-term shareholders supports our second hypothesis, H2.

Columns (3) and (4) of Table 3 present the results of Dissent voting. As the dependent variable in this table is a dummy variable, logistic regression is employed to conduct the analysis. Column (3) suggests that long-term institutional shareholders are less likely to cast dissent votes. Especially the increase in LTIO in a firm reduces the probability of having a proportion larger than 10% dissent votes of the total votes cast with regard to proposals suggested by a firm’s management. This finding is consistent with our results in Column (1) and with our first hypothesis H1. As we did with For voting, we examine the moderating effect of the cash deviation on the relationship between long-term institutional shareholders and Dissent voting. As can be seen from Column (4), long-term institutional shareholders increase their dissent votes when a firm holds excess cash. This might suggest that because long-term institutional shareholders closely monitor a firm’s cash policies, they are sensitive to any level of cash deviation from the optimal level. Our second hypothesis, H2, is supported by the results in Table 2.

Table 3, therefore, provides consistent results that are in line with our predictions in hypotheses H1 and H2. To sum up, the results in this table suggest that long-term institutional shareholders tend to voice their opinions by casting either positive or negative votes in regard to management proposals. On average, they tend to vote in favour of management recommendations, and their support is considerably larger when the firm maintains a low level of cash holding. However, long-term institutional shareholders show a different voting behaviour when the cash level increases relative to the optimal level. In particular, if a firm holds an excess level of cash, long-term institutional shareholders decrease their positive votes, and there is a higher probability of dissent voting exceeding 10% of the total votes cast.

To further ensure that long-term institutional shareholders change their vote in response to excess level of cash holdings, we re-estimate our main model to distinguish between proposals based on whether they are related to cash inflow or outflow. For example, proposals related to issuing shares and pre-emption rights are considered cash inflow proposals, and proposals related to approving/declaring dividends or share buybacks are considered cash outflow proposals.

Table 4. The voting of long-term institutional shareholders on cash inflow versus cash outflow proposals.

Dependent Variable =	For Voting		Dissent
Sample of Proposals	Inflow	Outflow	Inflow	Outflow
	(1)	(2)	(3)	(4)
LTIO	0.044 ***	0.022 ***	−2.136 ***	−3.185 **
	[0.012]	[0.007]	[0.639]	[1.353]
Excess Cash	0.023 **	0.010 *	−0.515	−0.602
	[0.009]	[0.006]	[0.327]	[0.690]
LTIO X Excess Cash	−0.046 ***	−0.018 *	1.553 **	2.379
	[0.017]	[0.010]	[0.657]	[1.713]
Constant	0.834 ***	0.811 ***	1.951	12.660 **
	[0.060]	[0.041]	[2.771]	[5.415]
Control Variables	Yes	Yes	Yes	Yes
Topic of proposal	Yes	Yes	Yes	Yes
Industry FE	Yes	Yes	Yes	Yes
Year FE	Yes	Yes	Yes	Yes
Adjusted_R2 (Pseudo_R2)	0.08	0.07	(0.13)	(0.22)
Obs.	5484	5124	5476	3556

This table presents the results of Pooled OLS and logistic regressions examining the moderating effect of excess cash on the relationship between the ownership of long-term institutional shareholders (LTIO) on positive and dissent votes while distinguishing between types of proposals. Columns (1) and (2) show the moderating effect of excess cash on the relationship between LTIO and (in favoure) voting. Columns (3) and (4) show the moderating effect of excess cash on the relationship between LTIO and dissent voting. Columns (1) and (3) use a subsample of management proposals that are related to cash inflow, and Columns (2) and (4) for proposals related to cash outflow. Detailed definitions of variables are reported in Table A1. Industry and year-fixed effects are used in all models. Standard errors are robust for heteroskedasticity (i.e., non-constant variance) and clustered by firm level to deal with protentional correlation within the firm. *, ** and *** indicate statistical significance of t-test at 10%, 5% and 1% level, respectively.

presents the results for these subsamples.

Columns (1) and (3) of Table 4 show the estimation results of our model using a subsample of proposals related to cash inflow, and Columns (2) and (4) for proposals related to cash outflow. Our focus here is on comparing the interaction terms of LTIO and excess cash between these subsamples. Columns (1) and (2) show that while long-term institutional shareholders decrease their positive voting as firms hold excess cash holdings on both subsamples, the decrease is more pronounced for the subsample of proposals related to cash inflow. Moreover, Columns (3) and (4) show that the moderating effect of excess cash on the relationship between long-term institutional ownership and dissent voting documented in Column (4) of Table 3 only exists for the subsample of proposals related to cash inflow, but not the subsamples of cash outflow proposals. The results in Table 4 provide supporting evidence for our hypotheses that long-term institutional shareholders vote in favour of management proposals, but they would decrease (increase) their positive (dissent) voting as firms hold excess cash, especially for proposals that are related to cash inflow.

4.3. Additional Tests

Additional tests are conducted to check the robustness of the reported findings. First, we use alternative measures for Dissent voting. Following Ertimur et al. (2011) and Stathopoulos and Voulgaris (2016a), we use a threshold of 20% instead of 10% to construct a Dissent voting variable, and we only use Against votes instead of both Abstain and Against votes to calculate Dissent voting.

Table 5. Alternative measures for shareholder dissent (robustness test).

Dependent Variable =	Dissent (20%)		Against (20%)
	(1)	(2)	(3)	(4)
LTIO	−2.429 ***	−3.589 ***	−2.133 **	−3.318 ***
	[0.853]	[1.082]	[1.038]	[1.266]
Excess Cash		−0.609		−0.552
		[0.466]		[0.536]
LTIO X Excess Cash		2.532 **		2.654 **
		[1.143]		[1.210]
Constant	4.185	4.933	7.477 *	8.116 *
	[3.842]	[3.879]	[4.084]	[4.215]
Control Variables	Yes	Yes	Yes	Yes
Topic of proposal	Yes	Yes	Yes	Yes
Industry FE	Yes	Yes	Yes	Yes
Year FE	Yes	Yes	Yes	Yes
Pseudo R-squared	0.17	0.18	0.22	0.22
Obs.	8153	8153	7746	7746

This table presents the results of Logistic regressions examining the effect of the ownership of long-term institutional shareholders (LTIO) on dissent (against and abstain) voting on proposals put to vote and the moderating effect of excess cash. Columns (1) and (3) replicate Column (3) in Table 3, and Columns (2) and (4) replicate Column (4) in Table 3 but use different measures of shareholder dissent. Dissent (20%) is a dummy variable that equals 1 if a proposal received dissent votes above 20% (where 10% was used in the main model) of the total votes cast and zero otherwise. Against is A dummy variable equals 1 if a proposal received only against votes (where both abstain and against votes were used in the main model) above 20% of the total votes cast, and zero otherwise. Detailed definitions of variables are reported in Table A1. Industry and year-fixed effects are used in all models. Standard errors are robust for heteroskedasticity (i.e., non-constant variance) and clustered by firm level to deal with protentional correlation within the firm. *, ** and *** indicate statistical significance of t-test at 10%, 5% and 1% level, respectively.

presents the results when using different measures of dissent voting, and the results are in line with the main findings in Table 3.

Moreover,

Table 6. Controlling for period effects (robustness test).

Panel A: Dependent Variable Is For Voting
Sample =	2000–2009	2010–2016	2000–2016
	(1)	(2)	(3)	(4)
LTIO	0.035 ***	0.029 **	0.032 ***	0.033 ***
	[0.012]	[0.011]	[0.008]	[0.008]
Excess Cash	0.009	0.019 **	0.016 **	0.016 **
	[0.010]	[0.009]	[0.006]	[0.006]
LTIO X Excess Cash	−0.018	−0.038 **	−0.032 ***	−0.032 ***
	[0.019]	[0.016]	[0.012]	[0.012]
annual_av_dissent			−0.249
			[0.155]
years2007_2009				0.001
				[0.002]
Constant	0.797 ***	0.847 ***	0.832 ***	0.828 ***
	[0.083]	[0.052]	[0.045]	[0.044]
Control Variables	Yes	Yes	Yes	Yes
Topic of proposal	Yes	Yes	Yes	Yes
Industry FE	Yes	Yes	Yes	Yes
Year FE	Yes	Yes	No	No
R-adjusted	0.07	0.12	0.08	0.08
Obs.	4292	5780	11,046	11,046
Panel B: Dependent variable is Dissent
Sample =	2000–2009	2010–2016	2000–2016
	(1)	(2)	(3)	(4)
LTIO	−2.633 **	−2.168 ***	−2.134 ***	−2.137 ***
	[1.041]	[0.747]	[0.594]	[0.600]
Excess Cash	0.288	−0.802 **	−0.445	−0.417
	[0.612]	[0.397]	[0.302]	[0.298]
LTIO X Excess Cash	0.047	1.939 **	1.427 **	1.399 **
	[1.402]	[0.792]	[0.617]	[0.614]
annual_av_dissenty			59.000 ***
			[14.039]
years2007_2009				−0.238
				[0.150]
Constant	4.826	3.540	1.989	3.152
	[4.496]	[3.198]	[2.495]	[2.485]
Control Variables	Yes	Yes	Yes	Yes
Topic of proposal	Yes	Yes	Yes	Yes
Industry FE	Yes	Yes	Yes	Yes
Year FE	Yes	Yes	No	No
Pseudo R-squared	0.20	0.19	0.18	0.17
Obs.	3429	5620	11,046	11,046

This table presents the results of OLS and Logistic regressions controlling for the financial crisis (2007–2009) period and the changing trend of shareholder activism. Columns (1) replicate the main model with only subsample periods from 2000 to 2009, and Column (2) replicate the main model with a subsample from 2010 to 2016. Columns (3) replicates the main model by controlling for annual average shareholder dissent. Column (4) replicates the main model with controlling for the financial crisis (2007–2009) period. The dependent variables are For (in favour) votes and Dissent votes in Panel (A) and (B), respectively. Detailed definitions of variables are reported in Table A1. Industry and year-fixed effects are used in all models. Standard errors are robust for heteroskedasticity (i.e., non-constant variance) and clustered by firm level to deal with protentional correlation within the firm. ** and *** indicate statistical significance of t-test at 5% and 1% level, respectively.

replicates our main models while controlling for the changing behaviour of shareholder voting and activism across different periods of time. The financial crisis of 2007–2009 raises many questions about the monitoring role of institutional investors, and prior studies document a change in voting behaviour and shareholder activism, which is attributed as a response to this crisis (e.g., McNulty and Nordberg 2016; Stathopoulos and Voulgaris 2016a; Tokbolat et al. 2019). Furthermore, prior research also shows that managers are more likely to expropriate shareholders through a corporate liquidity policy during a financial crisis (Tran 2020).

The results in Table 6 support the earlier evidence that shareholders became more active after the financial crisis. For instance, while the potential agency problem proxied by holding excess cash does not seem to affect the voting behaviour of shareholders before the financial crisis, this effect has changed after the crisis. The interaction between LTIO and excess cash is not statistically significant for the subsample period from 2000 to 2009 (Column 1) but is significant for the period after the crisis during 2010–2016 (Column 2). In Columns (3) and (4) of Table 6, instead of using subsample analysis, we replicate our main model while controlling for the annual average level of dissent voting (Column 3) and the period of the financial crisis (Column 4). Our results remain in line with the initial findings even after controlling for the changing behaviour of shareholder voting.

5. Discussion

This study explores the voting behaviour of long-term institutional shareholders, focusing on how their voting decisions are influenced by excess cash holdings, a key agency cost indicator. Our findings confirm that long-term institutional shareholders generally vote in favour of board recommendations, but this behaviour shifts when firms hold cash in excess of the optimal level, illustrating the influence of agency concerns.

Consistent with prior studies (Bushee 1998; Chen et al. 2007), we find that long-term institutional shareholders are more likely to support management proposals, reflecting their lower monitoring costs and ability to engage with firms over time (Stathopoulos and Voulgaris 2016a). This relationship aligns with previous work showing that increased long-term ownership reduces shareholder dissent, as these investors tend to trust and support management more (Derrien et al. 2013; Sauerwald et al. 2016).

Our key contribution is examining how excess cash moderates voting behaviour. We find that when firms hold cash in excess of the optimal level, long-term shareholders are more likely to vote against management recommendations, signalling their concerns about agency problems (Jensen 1986). This supports earlier research by Cleary and Wang (2017) and Huang and Petkevich (2016), who argued that institutional investors monitor and engage with firms to address excess cash. The negative voting behaviour is particularly strong when firms propose actions that could increase cash holdings, reflecting concerns about inefficiencies in cash management (Dittmar and Mahrt-Smith 2007).

Our analysis also reveals that long-term shareholders tend to dissent more on cash inflow proposals, such as share issuance, as these can exacerbate agency problems. In contrast, cash outflow proposals like dividends or buybacks receive more positive support, aligning with shareholders’ interests in reducing excess cash and improving governance (Sauerwald et al. 2016). This highlights the importance of cash policies in institutional investors’ voting behaviour (Orlova and Rao 2018; Yao and Hong 2023).

In line with the growing trend of shareholder activism post-financial crisis (Brav et al. 2024; Tokbolat et al. 2021), our findings suggest that long-term institutional shareholders have become more active in dissent voting. This shift reflects broader changes in shareholder engagement, where institutional investors use voting as a mechanism to address agency issues and improve corporate governance (McCahery et al. 2016; Sato and Takeda 2023).

6. Conclusions

This paper examines the voting behaviour of long-term institutional shareholders on management proposals during company general meetings and how such behaviour might change when considering the agency cost. This study focuses on UK public firms listed on the LSE during the period from 2000 to 2016. The UK provides an interesting context for studying the voting behaviour of shareholders due to the strong empowerment rights of shareholders (Davies et al. 2019) and the high proportion of institutional shareholders compared to other contexts.

In particular, we study the relationship between long-term institutional ownership and the positive voting on corporate capital proposals issued by a firm’s management; then, we test how this relationship is moderated by excess cash holdings as a proxy for the agency cost. We find a positive (negative) relationship between long-term institutional ownership and the voting in favour of (against) management proposals. We further show that this positive relationship is negatively moderated by the firms’ agency costs. This moderating effect only existed in the period after the recent financial crisis, which is consistent not only with the media view of the ‘Shareholder Spring’ after the crisis but also with evidence from prior studies on the change in voting behaviour by institutional shareholders as a response to the failure of many firms during the period (Tokbolat et al. 2019). Overall, these results provide evidence of the active role of long-term institutional shareholders and how they might use their voting rights as a channel to improve a firm’s corporate decisions.

This study offers valuable insights for institutional investors, corporate managers, and policymakers. For institutional shareholders, particularly long-term shareholders, the study emphasises the importance of actively monitoring firms’ cash holdings and using voting as a tool to address agency concerns. Corporate managers should ensure optimal cash management to avoid triggering shareholder dissatisfaction, and the board of directors should observe the voting outcome and behaviour to better understand shareholders’ preferences and evaluate firms’ decisions, policies and performance. For policymakers, the findings suggest that encouraging transparency in corporate governance and enhancing shareholder engagement mechanisms, especially around financial policies like cash holdings, could help mitigate agency problems and improve overall corporate governance.

Like all research, this study has several limitations that also present opportunities for future research. First, this study focuses on a UK sample, which may limit the generalizability of the findings to other countries with different corporate governance structures, regulatory environments, or shareholder behaviour. Second, the study relies on excess cash as a proxy for agency problems, but this measure may not fully capture the complexity of agency issues, as there could be other factors, such as managerial entrenchment or information asymmetry, that influence shareholder voting behaviour. Additionally, this study concentrates solely on voting as a mechanism of shareholder engagement, overlooking other potentially significant forms of institutional shareholder activism, such as direct dialogue with management, shareholder proposals, or public campaigns. These limitations suggest that future studies should broaden the sample to include firms from different countries, explore a more comprehensive set of agency proxies, and investigate multiple mechanisms of shareholder engagement to provide a more holistic understanding of institutional investor behaviour.