Previous studies show that a firm's relationship with its non-financial stakeholders influences a wide spectrum of corporate policies, such as capital structure choice (Bae et al., 2011, Banerjee et al., 2008, Kale and Shahrur, 2007, Maksimovic and Titman, 1991, Titman, 1984 and Titman and Wessels, 1988), corporate governance choice (Cen et al., 2012 and Johnson et al., 2012), the design of CEO compensation (Arora and Alam, 2005), information disclosure (Almazan et al., 2006), and earnings management (Raman and Shahrur, 2008). Although these studies enhance our understanding of the interaction between a firm's stakeholder relationships and corporate decisions, we know little about the impact of a firm's stakeholder relationships on its shareholders' welfare. In this paper, we develop this knowledge by investigating how a firm's customer–supplier relationships affect dividend payments to its shareholders.
A firm has customer–supplier relationships when its business depends on a small number of principal customers/suppliers. The extant literature suggests two possible channels through which a firm's customer–supplier relationships might have a negative impact on its dividend payments. On the one hand, developing the work of Banerjee et al. (2008), Kale and Shahrur (2007) and Titman (1984) argue that a firm with customer–supplier relationships often must undertake relationship-specific investments, which will, in turn, lead to higher financial distress costs. To cope with high financial distress costs, a firm must limit its use of debt and/or hold more liquid assets. Higher demand for liquid assets will reduce a firm's incentives to commit to paying dividends regularly. By choosing more flexible payout policies, a firm can save more liquid assets and improve its overall financial flexibility to mitigate the negative effect of high financial distress costs on firm value. We refer to this argument as the financial distress hypothesis.
On the other hand, Johnson et al. (2010b) indicate that by maintaining close product–market relationships, a principal customer has access to proprietary information concerning its supplier firms. The authors argue that a principal customer can serve as a monitoring and certifying entity for its suppliers, thereby reducing information asymmetry between the supplier firms and their shareholders. Consistent with this view, the study of Johnson et al. (2010a) documents an information spillover effect in the period of seasoned equity offering (SEO) announcements. Based on Jensen's (1986) classic free cash flow argument, the existing literature on payout policy (for example, John et al., 2011a, John et al., 2011b and Knyazeva, 2008) suggests that dividends are used to disburse free cash flow, thereby mitigating manager–shareholder conflicts. To the extent that a principal customer has incentives to choose and monitor financially healthy and well-qualified suppliers, its role as a monitoring and certifying entity can substitute for the effect of dividend payments in mitigating the free cash flow problem. We refer to this argument as the certification hypothesis.
Although both the financial distress hypothesis and the certification hypothesis predict a negative relationship between a firm's dependence on customer–supplier relationships and its dividend payments, the two arguments differ regarding the economic factor that causes this relation. The key difference is that the certification hypothesis does not involve financial distress. As long as 1) the suppliers' shareholders are concerned about the agency costs associated with free cash flow under information asymmetry, and 2) the shareholders believe that principal customers tend to select and monitor their suppliers carefully, the certification hypothesis should hold.
In this paper, we empirically investigate whether a firm's customer–supplier relationships affect its dividend policies. In our empirical tests, we use a firm's ratio of sales to principal customers over the total sales (hereafter referred to as BIGSALE) as a proxy for the importance of customer–supplier relationships to the firm's business. The information regarding sales to principal customers is provided by the Compustat business segment files. We use cash dividends over the market value of equity (hereafter referred to as DIV/ME) and cash dividends over net income (hereafter referred to as DIV/NI) as the measures of dividend levels. We find that over the period between 1981 and 2006, firms that rely more on customer–supplier relationships pay significantly lower dividends. This result holds true after controlling for a set of known determinants of dividend levels, such as firm size, growth opportunities, profitability, leverage, firm maturity, time-fixed effects, and industry-fixed effects or firm-fixed effects.
To determine further the nature of the negative relationship between a firm's dependence on customer–supplier relationships and dividend levels, we investigate how this relation is affected by relationship-specific investments. If this relation arises because a firm with customer–supplier relationships chooses a lower dividend level to mitigate the high financial distress costs associated with relationship-specific investments, then the financial distress hypothesis predicts that the negative relationship should be more pronounced for firms that are more likely to undertake relationship-specific investments. To the extent that relationship-specific investments facilitate a principal customer's access to proprietary information about its suppliers and increase its incentives to monitor them, the certification hypothesis predicts the same outcome. Consistent with expectations, our study reveals that the link between customer–supplier relationships and dividend level is more pronounced for firms that are more likely to undertake relationship-specific investments.
After documenting the cross-sectional relation between customer–supplier relationships and dividend levels, we examine the impact of customer–supplier relationships on a firm's dividend-smoothing behavior. Both the financial distress hypothesis and the certification hypothesis argue that the commitment to dividend payments is less important for firms with customer–supplier relationships, implying a negative relationship between dependence on customer–supplier relationships and dividend smoothing. We find that firms with customer–supplier relationships adjust their dividend levels significantly faster than those not in such relationships. This result suggests that a firm with customer–supplier relationships exhibits less dividend smoothing; thus, the result supports the prediction of the two hypotheses. In addition, we examine the effect of customer–supplier relationships on dividend changes. If a firm with customer–supplier relationships is less committed to regular dividend payments, the firm would increase dividends more conservatively and cut dividends more aggressively. Consistent with this conjecture, our results show that a firm relying more on customer–supplier relationships is less likely to increase dividends and more likely to cut dividends. Moreover, conditional on the decision to change dividend levels, the magnitude of dividend changes is negatively related to a firm's dependence on customer–supplier relationships.
To verify further the robustness of our results, we perform the following two tests. First, we conduct an out-of-sample test using information regarding a firm's purchases from its dependent suppliers. By identifying Compustat firms that are reported as principal customers by their dependent suppliers, we construct the ratio of purchases from dependent suppliers to the costs of goods sold (hereafter referred to as BIGBUY). Because customer–supplier relationships are more important to a firm that purchases a higher proportion of its input from dependent suppliers, we expect such a firm to pay lower dividends. Our results are robust to this alternative measure of the importance of customer–supplier relationships to a firm's business.
Second, we examine a firm's dividend payments at the three stages of its customer–supplier relationships: the pre-relationship period, the establishment of the first customer–supplier relationship, and the dissolution of the last customer–supplier relationship. We find that a firm in the pre-relationship period does not pay lower dividends than peer firms that have no customer–supplier relationships during the entire sample period, which mitigates the concern that unobservable firm characteristics may drive the negative relationship between the importance of customer–supplier relationships and dividend payments. Moreover, we find a decrease in dividend levels after a firm establishes its first customer–supplier relationship and an increase in dividend levels after the firm dissolves the customer–supplier relationship with its last principal customer. These results provide further evidence that a firm's dependence on customer–supplier relationships has a negative effect on its dividend payments.
Given our strong evidence regarding the negative relationship between a firm's dependence on customer–supplier relationships and its dividend payments, we conduct several tests to evaluate the extent to which the two hypotheses (the financial distress hypothesis versus the certification hypothesis) can explain this negative relationship. To evaluate the financial distress hypothesis, we examine how this negative relationship differs across firms with different likelihoods of entering financial distress. Consistent with the financial distress hypothesis, our results show that the negative relationship is more pronounced for firms that are more likely to enter financial distress.
For the certification hypothesis, we examine how the free cash flow problem affects the negative relationship between a firm's dependence on customer–supplier relationships and its dividend payments. The certification hypothesis predicts that this negative relationship is more pronounced for firms with more severe free cash flow problems. Our results do not support this prediction.
Finally, we examine how a firm's dependence on principal customers is related to abnormal stock returns upon a dividend cut announcement. To the extent that a dividend cut signals increased probability of financial distress, an announcement by a firm with customer–supplier relationships would cause a more negative stock return because such a firm would suffer more from financial distress. Thus, the financial distress hypothesis predicts that an announcement return is negatively related to a firm's dependence on customer–supplier relationships. By contrast, if the certification effect is the dominant factor in the link between customer–supplier relationships and dividend payments, the existence of customer–supplier relationships would mitigate the negative signal conveyed by the dividend cut announcement. Therefore, the certification hypothesis predicts a positive relationship between the importance of customer–supplier relationships and announcement returns. Our results again support the financial distress hypothesis but do not support the certification hypothesis.
As the first empirical study that investigates the effect of a firm's customer–supplier relationships on its shareholders' welfare, this paper contributes to the literature on how a firm's relationship with its non-financial stakeholders affects its corporate decisions. We provide evidence that a firm's relationship with its customers/suppliers is an important determinant of its shareholders' income. Moreover, our results suggest that concern regarding financial distress is the key channel through which a firm's customer–supplier relationship affects its shareholders' income.
The paper is closely related to the literature on the effect of a firm's customer–supplier relationships on its capital structure choices. Titman (1984) indicates that stakeholders' incentives to conduct relationship-specific investments affect a firm's financing decisions. He argues that because stakeholders face switching costs if a firm is liquidated, their incentives to undertake relationship-specific investments depend on a firm's financial condition. Developing Titman's (1984) model, Banerjee et al. (2008) and Kale and Shahrur (2007) show that when relationship-specific investments are important to a firm's business, the existence of customer–supplier relationships has a negative effect on the firm's leverage ratios. Overall, these studies provide evidence that relationship-specific investments affect a firm's external financing decisions. In contrast, our paper examines whether and through which channels a firm's dependence on major customer–supplier relationships affects its shareholders' income.
Additionally, our paper is related to recent studies on the information effect of principal customers. Johnson et al. (2010b) find that initial public offering (IPO) firms that have product market relationships with large customers experience higher valuation and better long-term performance. The authors interpret their results as evidence that a principal customer plays a certifying and monitoring role and thereby reduces information asymmetry between the suppliers and their shareholders. Johnson et al. (2010a) examine dependent suppliers' SEOs and document an information spillover effect down the supply chain. The authors provide further evidence that entities on the supply chain are important sources of information to investors. Although we do not find evidence supporting the certification hypothesis in the context of dividend policies, our results are not necessarily inconsistent with the above findings because investors' concern regarding information asymmetry is probably a much more important factor in a firm's IPO/SEO decisions than in its dividend policies.
The paper is organized as follows. In Section 2, we describe the data used in this paper and provide summary statistics. In Section 3, we examine the cross-sectional relation between the importance of customer–supplier relationships and dividend level. In Section 4, we investigate how customer–supplier relationships affect dividend smoothing and dividend changes. In Section 5, we perform several robustness tests. In Section 6, we evaluate the extent to which the financial distress hypothesis and the certification hypothesis can explain the negative relationship between the importance of customer–supplier relationships and dividend level. Section 7 summarizes and concludes the paper.
