Are dividends and share repurchases substitutes? : evidence from the reaction of financial analysts

In this study, I propose a novel test that can contribute towards resolving whether repurchases do in fact substitute for dividends. I contend that using the reactions of financial analysts to announcements of dividends initiations / increases and share repurchases can effectively test whether repurchases do substitute for dividends. If the substitution hypothesis holds, then the analysts' changes in recommendations following the announcement of a dividend initiation or increase and a repurchase announcement should be the same, ceteris paribus. My findings show that when analysts' reactions include no changes, positive and negative changes in recommendation, dividends and repurchases are viewed as substitutes. However, when there is an actual change in the analysts' recommendation following a distribution announcement, the analysts downgrade more and upgrade less the firms that announce share repurchases programs as opposed to dividend initiations / increases. They are also more likely to make a recommendation change when there is a share repurchases announcement. Further, I explore whether the analysts' preference might be justified. Specifically, I investigate how the signal quality of the payout method, the potential for earnings management of the payout method, or behavioral biases might play a role in the analysts' decision making process. I find that, given an actual change in recommendation, financial analysts upgrade less and downgrade more firms with share repurchases announcements when the firms are associated with higher earnings management, during period of higher demand for dividends, and when the firms have been downgraded in the month preceding the announcement. Financial analysts view dividends and repurchases as substitutes post 2004 and when repurchases are classified as serial repurchases.