Do firms use share repurchases as a way to manipulate stock prices?

本文主要研究美國上市公司宣告買回庫藏股的動機。相較過去的文獻，我們提出一個新的觀點，認為公司可能利用宣告買回庫藏股當作是操弄股價的一種手段。主要研究發現，如果買回庫藏股公司的裁量性會計應計項目偏高，該公司就不能獲得正的長期股票報酬。這表示公司可能在操控盈餘失敗之後，轉而利用宣告買回庫藏股來做為操控股價的另一種手段。我們驗證這一類型公司與過去的理論是否有關，結果發現傳統的各項理論，如釋放公司價值被低估的訊息，分配自由現金流量，調整資本結構，或是準備員工認股選擇權等理論，都無法解釋裁量性會計應計項目偏高的公司為何要宣告買回庫藏股。其他面向的研究，如長期的公司營運績效，公司季報宣告效果，分析師的盈餘預測以及公司實際購回庫藏股的分析，也都支持我們的假說。所以我們相信，除了傳統的各項動機，公司也有可能利用宣告買回庫藏股，當成操弄股價的一種工具。This paper investigates the motivation of open market share repurchases. A new aspect of the stock price manipulation is provided to shed lights on the source of open market share repurchases. We suggest repurchases with frequent high discretionary accruals being motivated by the stock price manipulation. Repurchase firms with high discretionary accruals tend to announce repurchases to boost up the stock prices. However, it mainly argues that repurchase firms with high discretionary accruals earn insignificant returns in the long run, which is against the undervaluation hypothesis. Other examinations such as operating performance, earnings surprise, analysts forecast and other results suggest that repurchase firms with high discretionary accruals are not motivated by traditional hypotheses. Therefore, these firms are likely to announce repurchases to boost up the stock price rather than other positive reasons.1.Introduction---3 2.Literature reviews---6 3.Data and preliminary results---9 3.1 Data---9 3.2 Accruals and discretionary accruals---10 4.Long run performance of DA quintiles---13 4.1 Long run stock return---13 4.1.1 Long run buy-and-hold returns---13 4.1.2 Carhart 4-factor model---16 4.2 Long run operating performance---20 4.2.1 Long run abnormal operating performance---20 4.2.2 Other measures of operating performance and earnings components---24 5.Test of other hypotheses---26 6.Reaction from the market and analysts---28 6.1 Quarterly earnings announcement returns---28 6.2 Revision of analysts EPS forecast---30 7.Actual buyback activities, regression result and some robustness---33 7.1 Actual buyback and manipulation hypothesis---33 7.2 Some robustness---36 8.Conclusion---3