Does the information content of payout initiations and omissions influence firm risks?

Abstract

We study the influence on firm risks of NASDAQ and NYSE firm payout initiations and omissions. These payout events can be interpreted as managerial signals of firm financial life-cycle maturation resulting in concomitant changes in firm risks. We remove confounding payout types and we match on the propensity to initiate or omit informed by determinants of payout known to investors in advance. For payout event and matched firms, we apply the difference-in-differences method to estimate the effect of the information content of actual initiations and omissions on firm risks. We find consistent significant declines in total, aggregate systematic, and idiosyncratic firm risks after cash dividend initiations and increases after dividend omissions, but only incidentally after share repurchase initiations and omissions.

Introduction

Payout policy and firm risks are two major topics in the field of corporate finance, but the influence of extreme payout events (payout initiations and omissions) on firm risks has received scant attention in the literature. Notwithstanding, Brav et al. (2005) report that many financial executives believe that there is a causal relation between payouts and risk changes. The objective of this paper is to test whether changes in firm risks are indeed attributable to the new information content provided to investors in payout initiations and omissions. We assess the risk effects of dividends and repurchases, of initiations and omissions, on total, aggregate systematic, and idiosyncratic risks. The findings in respect to the estimated risk effects are not only relevant in their own right, but also because these risk effects are very likely to influence firm values.

In line with the semi-strong form of the efficient market hypothesis (Fama, 1970) investors can have expectations, based on publicly available information, on prospective payout events, which are reflected in price behaviour. However, as managers can have more information than investors, firms with a similar likelihood of a payout event, from the perspective of investors, do not necessarily exhibit the same incidence of actual payout events. Payout (especially dividend) initiation and omission events, which are conducted by management, reveal new information to investors, which results in positive value effects after initiations and negative value effects after omissions (Christie, 1994, Ikenberry et al., 1995, Charitou et al., 2011, Bonaimé, 2012). Moreover, the value effects based on the new information are smaller for initiations than for omissions (Michaely et al., 1995).

The recent financial risk related literature suggests that the value effects associated with payouts can be due to the new information content on firm maturity signalled to investors in the payout. A dividend initiation may indicate that a firm transits from a growth to a maturity phase, and that it has fewer growth options relative to assets in place. The new information signalled to investors results in lower systematic risks (Berk et al., 1999), and concomitant lower discount rates and increases in the firm value (Grullon et al., 2002, Grullon and Michaely, 2004). We add to the maturity hypothesis that the value effects of the information content of the payout signal can also arise from changes in firms’ idiosyncratic risk. As idiosyncratic risk is indicative of growth opportunities (Hoberg and Prabhala, 2009) and cash flow risk (Bulan et al., 2007), it is also expected to decline in the transition from a firm’s high growth to a lower growth phase. For this reason we investigate whether the information signals provided by payout initiations also influence idiosyncratic firm risk.

In principle, risk related value effects, moreover, do not only arise from signals on firm maturation after payout initiations. If a firm faces improved investment opportunities, which can be interpreted as a return to an earlier phase in its financial life-cycle, it may omit payouts if it finds it optimal to finance the new opportunity internally. Concomitantly, the firm’s systematic and idiosyncratic risks may increase and its value may decline. Of course, a firm may also omit if it tries to cope with pure financial distress, and that may also signal to investors increased firm risks and a reduction in the firm value.

As we wish to address the risk effects of extreme payout events comprehensively, we distinguish dividend and share repurchase payout channel events. It is likely that there are different anticipated effects on firm risks according to the payout channel involved. There are different stylised contexts in which dividend and repurchase payout policies are realised and dividends and repurchases are not perfect substitutes (Bhargava, 2010). When a firm progresses in its financial life cycle, from a transitory income to a predominantly permanent income (Jagannathan et al., 2000), a firm commences payout using repurchases, and with the maturation of the firm’s income stream, it may decide to initiate paying dividends (Grullon et al., 2002, Grullon and Michaely, 2004). In comparison to unexpected repurchase initiations, unexpected dividend initiations may thus contain a stronger signal of maturation to investors and have larger risk effects.

We apply propensity scores to identify comparable firms (Rosenbaum and Rubin, 1983) and non-parametric local linear regression matching to facilitate difference-in-differences tests (Heckman et al., 1998a, Heckman et al., 1998b, Guo and Fraser, 2010). Contrary to the extant literature which matches on risk factors, we utilise a matching procedure for the selection of the counterfactual firms that is based on the publicly available information in the capital market. Our counterfactual firms therefore have a comparable, ex ante, propensity to initiate (or omit) payouts as the firms which actually initiate (or omit). In respect to the resulting firms, investors are unable to distinguish whether they will initiate (or omit). We compare the risk changes of firms that do initiate (or omit) to those of the firms that do not signal such additional information to the market. In addition, our methodology corrects for confounding payout channels.

The paper is organised as follows. Section  2 states the tested hypotheses and explains why we use a difference-in-differences propensity score matching (PSM) methodology. Section  3 presents information about the dataset and associative descriptive statistics. Section  4 presents the main results concerning “causal” relations between the information content of payout initiations and omissions and firm risks. Section  5 reports the conclusions.