2.1 Institutional background

Stakeholder constituency statutes were introduced in response to ongoing discussions in corporate law and finance over the extent to which firms should serve broader societal interests beyond shareholders (e.g., Bainbridge, 1991; Orts, 1992). This discourse can be traced back to the early 20th century, particularly to Dodd’s (1931) influential argument that corporations ought to fulfill public responsibilities and not focus solely on maximizing shareholder wealth. In contrast, many economists and legal scholars maintain that shareholder primacy is essential for effective corporate control, arguing that the interests of other stakeholders are already protected through existing legal contracts and regulatory frameworks (Tirole, 2001; Bénabou and Tirole, 2010).

This ideological divide regained prominence in the 1980s as stakeholder theory gained traction and U.S. states began to adopt statutes allowing corporate directors to explicitly consider non-shareholder constituencies. These statutes were introduced during an era of heightened takeover activity and were intended, in part, to provide boards with legal cover to resist hostile bids (Karpoff and Wittry, 2018). While stakeholder constituency statutes do not prescribe a specific hierarchy among stakeholder interests, they permit directors to legally consider employees, customers, and other groups when making decisions, even when such considerations may not align with the goal of maximizing shareholder value in the short term (Bainbridge, 1991; Lee, 2022).

As of 2007, 35 U.S. states had implemented stakeholder constituency statutes. In line with Flammer and Kacperczyk (2016), we provide a summary of the adoption timeline and identify the corresponding firms affected, as shown in Table 2. These legal reforms offer a unique empirical setting for identifying shifts in corporate priorities, as they alter the legal basis of boardroom decision-making by broadening the range of interest directors may consider. Importantly, the enactment of these statutes is determined at the state level and is exogenous to any individual firm’s corporate strategy, allowing us to treat their adoption as plausibly exogenous variation in stakeholder orientation.

While stakeholder-oriented strategies have been associated with long-term advantages, such as greater innovation and improved employee involvement (e.g., Flammer and Kacperczyk, 2016; Chowdhury et al., 2021), they may not always receive full support from shareholders. Ni (2020) emphasizes that shareholders might resist stakeholder-focused reforms due to concerns over short-term financial sacrifices, such as reduced dividend payouts or lower accounting transparency. This skepticism is rooted in the perception that directors could invoke stakeholder claims to justify decisions that dilute shareholder value¹.

Survey evidence also indicates that corporate directors themselves often view their roles as encompassing responsibilities to a broad range of constituents. For example, Wang and Dewhirst (1992) find that directors of large U.S. public firms acknowledge the importance of addressing stakeholder concerns in their governance roles. Taken together, these findings suggest that although stakeholder constituency statutes do not impose mandatory duties beyond shareholders, their permissive nature can meaningfully reshape managerial priorities and serve as a legal foundation for moving away from exclusive shareholder primacy (Flammer and Kacperczyk, 2016).

2.2 Prior literature and hypothesis development

Dividend smoothing, defined as the practice of maintaining stable and predictable dividend payouts despite earnings fluctuations, has long been viewed as a central feature of corporate financial policy since Lintner’s (1956) foundational work. Among the various theories proposed to explain this phenomenon, the agency-based perspective has gained prominence for its emphasis on the role of dividend policy in mitigating managerial discretion and aligning the interests of managers and shareholders².

From the agency perspective, smoothing dividends can function as a commitment device that restricts the availability of free cash flow under managerial control, thereby reducing the scope for opportunistic behavior. By maintaining a stable payout level, managers impose self-discipline and subject the firm to the scrutiny of external capital markets, especially when additional financing is needed (Easterbrook, 1984; Jensen, 1986). Regular scrutiny from external financial markets serves as a powerful check on managerial behavior, ultimately curbing agency costs (Leary and Michaely, 2011). In addition, Allen et al. (2000) and Shin and Kim (2014) document that consistent dividend payouts can enhance a firm’s appeal to institutional investors, who value predictable income streams and actively engage in monitoring management. These investors are sensitive to dividend cuts and can impose significant reputational and financial penalties, encouraging managers to avoid erratic dividend behavior.

Recent studies provide empirical support for this view. For example, Leary and Michaely (2011) find that firms facing greater external scrutiny, such as those with higher analyst coverage, lower earnings volatility, and larger firm size, are more likely to engage in dividend smoothing. Similarly, Javakhadze et al. (2014) show that companies operating in environments with weaker investor protection tend to smooth dividends more, suggesting that dividends can serve as informal governance mechanisms under such conditions. In a related context, Li et al. (2023) investigate the impact of China’s Green Credit Policy, which imposed tighter financing constraints and enhanced external monitoring on polluting firms. They find that firms subject to the policy experienced a significant reduction in dividend smoothing, indicating that increased external oversight, such as stricter credit regulation, can substitute for dividend smoothing in mitigating agency conflicts.

In this context, the enactment of stakeholder constituency statutes can alter the degree to which firms depend on dividend smoothing to address agency problems. By granting directors greater latitude to consider the interests of non-shareholder groups, such statutes may affect the perceived need for governance mechanisms, including stable dividend payouts. However, the direction of this influence remains theoretically ambiguous.

On one hand, stakeholder orientation can reduce agency concerns by encouraging managers to act with stronger ethical and social responsibility. According to the ethical perspective of stakeholder theory (Jones, 1995; Donaldson and Preston, 1995), companies that take into account the interests of a broad range of stakeholders, including shareholders and non-shareholder stakeholders, are more likely to behave ethically. This kind of ethical behavior helps build trust and long-term cooperation between the firm and its stakeholders, which can make the firm more stable over time. Although stakeholder orientation is not exactly the same as corporate social responsibility (CSR), it can still make managers more aware of the broader social impact of their decisions. When managers consider how their actions affect employees, customers, and the community, they may place greater value on acting responsibly and ethically. Prior evidence suggests that higher ethical standards enhance the credibility of managerial actions and reduce concerns about opportunistic behavior (Kim et al., 2025). As a result, shareholders may become less concerned about potential misuse of free cash flow, which lowers the need for dividend smoothing as a way to control managerial behavior (Ni, 2020; Chowdhury et al., 2021; Gao et al., 2021).

Based on this mechanism, our first hypothesis is as follows:

Hypothesis 1a (H1a): Stakeholder orientation reduces firms’ dividend smoothing practices.

On the other hand, giving more attention to stakeholder interests could create more chances for managers to act in their own self-interest. For example, they might misuse company resources or make deals that benefit employees or other non-shareholder groups instead of shareholders. From this perspective, stakeholder orientation may raise concerns that managers could collude with non-shareholder stakeholders, such as suppliers or employees, to use corporate resources in ways that benefit those parties rather than shareholders (Bertrand and Mullainathan, 2003; Pagano and Volpin, 2005). Shareholders may worry that such collusion allows managers to divert free cash flow toward stakeholder interests under the guise of stakeholder engagement. In response to these concerns, managers might smooth dividends more actively to reassure investors that they remain committed to disciplined financial practices and the protection of shareholder value.

This alternative mechanism leads to the following hypothesis:

Hypothesis 1b (H1b): Stakeholder orientation increases firms’ dividend smoothing practices.

3.1 Measurement of dividend smoothing

Following prior literature (e.g., Leary and Michaely, 2011; Brockman et al., 2022), we construct two measures of dividend smoothing. Both measures rely on the speed of adjustment, estimated through a two-stage procedure, and differ only in the length of the estimation window used to construct the smoothing variable. In the first stage, we compute a firm’s target payout ratio ($TPR_i$) using a rolling window approach. Specifically, we calculate the median of the firm’s historical payout ratios over a trailing 5-year or 10-year window ending in year $t$, requiring at least three or seven valid observations and a minimum of one positive dividend payout for each calculation period. We define the payout ratio as the proportion of dividends to income before extraordinary items. We then calculate the deviation between this target and the actual payout in year $t$, denoted as $dev_{i,t}$.

In the second stage, we again rely on a rolling window regression to estimate the adjustment coefficient $\beta$ by regressing changes in dividends per share ($\Delta DPS$) on the deviations ($dev_{i,t}$) from the target payout ratio.

We calculate dividends per share ($DPS$) as total dividends divided by the number of common shares outstanding, and earnings per share ($EPS$) as income before extraordinary items divided by the same denominator. Both $DPS$ and $EPS$ are adjusted for stock splits using the appropriate adjustment factors. The speed of adjustment, denoted as $SOA5$ or $SOA10$, reflects a firm’s responsiveness to past deviations, based on a trailing 5-year or 10-year window, respectively. Lower values of $SOA5$ or $SOA10$ indicate stronger dividend smoothing behavior. To mitigate the influence of outliers, we truncate the distribution of the speed of adjustment estimates at the 2.5 and 97.5 percentiles. This approach ensures comparability with prior studies (e.g., Leary and Michaely, 2011; Larkin et al., 2017; Brockman et al., 2022) that adopt the same cutoffs.

3.2 Research design

To test the impact of stakeholder orientation on dividend smoothing, we rely on prior studies and estimate the following difference-in-differences model:

where we use the speed of adjustment (denoted as SOA5 or SOA10) as the dependent variable. We define CS as an indicator for stakeholder orientation, taking the value of one for firm-years in which the state of incorporation has adopted stakeholder constituency statutes, and zero otherwise.

To isolate the effect of stakeholder orientation, we follow prior studies (e.g., Leary and Michaely, 2011; Larkin et al., 2017) and control for a wide range of firm-level attributes that could independently affect dividend smoothing. Firm size (SIZE), sales growth (GROWTH), and profitability (ROA) capture a firm’s financial capacity and growth opportunities, while firm age (AGE) and the frequency of reported losses (PER_LOSS) reflect organizational maturity and historical performance consistency. The book-to-market ratio (BM) and leverage (LEV) account for firm valuation and financial risk, respectively, which can affect dividend policy through capital market discipline. We also control for earnings volatility (EARN_VOL) and cash holdings (CASH) to capture uncertainty and liquidity, which are directly related to a firm’s ability and need to smooth dividends. R&D intensity (RD) and capital expenditures (DCPX) proxy for internal investment demands that may compete with dividend payouts. Lastly, we include dividend yield (DVY), share turnover (TURNOVER), and analyst following (AF) to reflect the firm’s capital market environment and the degree of external monitoring. In addition, we control for unobserved heterogeneity by including firm fixed effect ($f_i$) fiscal year fixed effects ($r_t$) and state of headquarter location fixed effects ($\gamma_s$). The fixed effects help account for time-invariant firm characteristics, aggregate year-specific shocks, and persistent differences across states, respectively. Since stakeholder orientation is determined at the incorporation-state level, we cluster standard errors accordingly. This accounts for within-state serial correlation in unobserved factors that may jointly influence firms’ dividend behavior over time (Bertrand and Mullainathan, 2003; Bertrand et al., 2004).

3.3 Sample selection

Table 1

4.1 Descriptive statistics and correlations

Table 2

A notable clustering of adoptions occurred during the late 1980s and early 1990s, with states such as New York (1987), Ohio (1984), and Pennsylvania (1990) enacting these statutes and covering 81, 67, and 47 firms, respectively. Conversely, several states, including New Mexico (1987), Idaho (1988), Kentucky (1989), North Dakota (1993), and Vermont (1998), report no affected firms in the sample, likely due to minimal incorporation activity or limited application of the statutes within those jurisdictions.

Panel A of Table 3

Several states are excluded from the table because no firms in the final sample were affected by their adoption of stakeholder constituency statutes. These states include New Mexico (1987), Idaho (1988), Kentucky (1989), North Dakota (1993), and Vermont (1998).

4.2 The impact of stakeholder orientation on dividend smoothing

Table 4

With respect to the control variables, SIZE is negatively associated with the speed of adjustment, consistent with the notion that larger firms face less information asymmetry and thus have weaker incentives to smooth dividends. ROA is positively associated with the speed of adjustment, suggesting that more profitable firms tend to smooth dividends to a lesser extent. We also find that firm age (AGE) is negatively associated with the speed of adjustment. This result indicates that older firms smooth dividends more strongly, consistent with the notion that mature firms prioritize dividend stability as a means of maintaining their reputation and meeting long-term shareholder expectations. In contrast, firms with persistent losses (PER_LOSS) exhibit significantly less dividend smoothing. Firms under prolonged financial distress are less able to sustain stable dividend policies. Instead, they tend to adjust payouts more rapidly in response to earnings shortfalls. Our findings for the control variables are broadly consistent with the patterns documented in earlier studies (e.g., Leary and Michaely, 2011).

Overall, the results reported in Table 4 indicate that the adoption of stakeholder constituency statutes is associated with a statistically and economically significant decline in dividend smoothing. This result is consistent with the hypothesis (H1a) that stakeholder orientation alters managerial incentives by mitigating agency concerns between shareholders and managers. When managers are granted broader discretion to consider stakeholder interests, the need to rely on dividend smoothing as a mechanism to reduce agency problems may diminish.

To investigate this, we estimate an event-study specification by replacing the CS indicator in Equation (2) with a series of leads and lags relative to the year of statute adoption: $CS^{-4}$, $CS^{-3}$, $CS^{-2}$, $CS^{-1}$, $CS^{0}$, $CS^{+1}$, $CS^{+2}$, $CS^{+3}$, $CS^{+4}$, $CS^{+5}$, and $CS^{6+}$. Each indicator captures whether a firm-year falls within a specific time window before or after the enactment of the statute in the firm’s state of incorporation. For instance, $CS^{-4}$ ($CS^{+1}$) equals one if the observation occurs four years before (one year after) the law’s adoption, and zero otherwise.

4.3 Dynamic effects

This section evaluates whether the parallel trends assumption holds and explores potential endogeneity issues, with a particular focus on the risk of reverse causality. Specifically, we examine whether changes in dividend smoothing precede the adoption of stakeholder constituency statutes. As reported in columns (1) and (2) of Table 5

Figure 1

4.5.1 Stakeholder salience and dividend smoothing

To further explore the underlying mechanisms linking stakeholder orientation to dividend smoothing, we conduct subsample analyses based on firm characteristics that heighten the salience of stakeholder interests: industry consumer orientation, labor intensity, and liquidation value. In addition to capturing contexts in which stakeholder interests are more salient, these three characteristics also reflect environments where potential conflicts between shareholders and non-shareholder stakeholders are more likely to arise. By examining variation along these dimensions, we assess whether the impact of stakeholder constituency statutes is amplified when stakeholders hold relatively weaker bargaining positions or when their interests conflict more directly with those of shareholders.

We begin by examining whether the observed reduction in dividend smoothing following the adoption of stakeholder constituency statutes is more pronounced among firms in industries where customer interests are particularly salient. Prior research suggests that stakeholder-oriented initiatives can strengthen a firm’s reputation and foster customer loyalty by signaling ethical behavior and social responsibility (Brown and Dacin, 1997; Luo and Bhattacharya, 2006; Porter and Kramer, 2006). These effects are likely to be more pronounced in industries that interact directly with end consumers, where reputational concerns are more salient. Flammer and Kacperczyk (2016) provide empirical support, showing that stakeholder orientation has stronger innovation effects in consumer-facing sectors. Based on this logic, we expect the effect of stakeholder constituency statutes adoption on dividend smoothing to be more pronounced in consumer-focused industries, where enhanced stakeholder legitimacy may reduce shareholder pressure and weaken the need for dividend smoothing as a tool to mitigate agency conflict.

To test this prediction, we construct a binary variable, CONSUMER, which equals one for firms operating in industries classified as “consumer goods,” and zero otherwise, following Lev et al. (2010)¹². We then estimate our baseline model separately for firms in consumer-goods industries and firms outside these industries to assess whether the impact of stakeholder orientation varies across these industry segments. Panel A

Next, we examine whether the effect of stakeholder constituency statutes on dividend smoothing varies by labor intensity. Following Gao et al. (2021), we define labor intensity as the number of employees scaled by sales revenue, reflecting how deeply a firm’s operations depend on its workforce. When stakeholder constituency statutes are adopted, managers gain formal discretion to consider employee interests (Garman and Kubick, 2024). In labor-intensive firms, this shift in legal framework enables managers to reallocate free cash flow to benefit employees, such as training, retention programs, or enhanced workplace conditions, rather than constraining it through smoothed dividends. Accordingly, we predict that firms with high labor intensity are more likely to benefit from the governance shift enabled by stakeholder constituency statutes.

To test this prediction, we estimate our baseline regression model separately for firms with high and low labor intensity. Panel B reports the regression results. In columns (1) and (3), we find that the coefficient on CS is significantly positive in the high-labor-intensity subsample. In contrast, the coefficient on CS is statistically insignificant in the low-labor-intensity subsample. These results suggest that the reduction in dividend smoothing behavior is primarily driven by firms where employees play a more central operational role.

Lastly, we explore whether the effect of stakeholder constituency statutes on dividend smoothing varies with firms’ liquidation value, a proxy for the relative bargaining power of creditors. Acharya et al. (2007) argue that when a firm has a low liquidation value, equity holders possess stronger bargaining leverage relative to creditors, as they can credibly threaten to leave creditors with limited recovery in the event of default. In such environments, stakeholders, particularly debtholders, face weaker protection, and shareholder dominance tends to be more pronounced in the absence of formal safeguards.

Following Berger et al. (1996), we compute liquidation value as the estimated recoverable value of firm assets and classify firms as having low or high liquidation value based on the sample median.¹³ We then estimate our baseline regression separately for these two subsamples. Panel C presents the corresponding results. In columns (2) and (4), we observe a statistically significant positive association between CS and the speed of adjustment in the low-liquidation-value group. By contrast, in the high-liquidation-value group, the coefficients on CS are small and statistically insignificant. These results suggest that stakeholder constituency statutes have a stronger impact in settings where stakeholders are initially in a weaker position to influence managerial behavior.

Overall, the results reported in Table 7 indicate that the effect of stakeholder constituency statutes on dividend smoothing is more pronounced in settings where stakeholder concerns are more salient and where the potential conflict between shareholders and non-shareholder stakeholders is stronger.¹⁴

4.5.2 Agency costs and dividend smoothing

To complement the cross-sectional analysis based on stakeholder salience, we also examine whether the effect of stakeholder constituency statutes varies with firms’ agency costs. We predict that when managers have greater access to discretionary internal funds, the scope for opportunistic behavior is larger, and decisions regarding payout smoothing are more sensitive to governance constraints. Accordingly, if stakeholder constituency statutes function as a governance mechanism by limiting managers’ opportunistic behavior, their impact should be more pronounced among firms with higher agency costs.

We capture agency costs using two standard proxies that reflect the extent of internal resources available to managers: cash holdings (CASH) and free cash flow (FCF), measured following Javakhadze et al. (2014). CASH is defined as cash and short-term investments scaled by total assets, and FCF is computed as operating cash flow minus cash dividends, scaled by total assets. Based on the median values of these two proxies, we classify firms into high- and low-agency-cost groups and re-estimate our baseline model (i.e., equation (2)) for each subsample.

Table 8 reports the results. In Panel A, the CS coefficient is positive and statistically significant only for high-cash firms when using both SOA5 and SOA10, whereas the estimate for low-cash firms are smaller and statistically insignificant. A similar pattern emerges with the free-cash-flow subsample in Panel B. CS is significant in the high-FCF group across both smoothing measures, but insignificant in the low-FCF group. These results suggest that the reduction in dividend smoothing following the constituency statute adoption is more pronounced among firms with higher agency costs.

Taken together, the evidence from Table 7 and Table 8 indicates that stakeholder constituency statutes reduce dividend smoothing most strongly when stakeholder interests are more salient and when managerial discretion is greater. These results suggest that the influence of the stakeholder constituency statutes is affected by the prominence of stakeholder claims and the extent of internal resources available to managers, consistent with an agency-based interpretation of firms’ payout decisions.