Insider ownership and financial flexibility

Abstract

This study examines the cash flow sensitivity of external financing (financial flexibility (FF)) of firms from the perspective of the agency by focusing on whether and how insider ownership has a modifying effect on the FF of firms. Given the documented relationship between corporate governance and cost of external financing, insider ownership, an important proxy of corporate governance, should affect FF via the agency cost channel. A review of the US nonfinancial firms from 1992 to 2009 indicates that insider ownership aids in determining the FF of firms. More specifically, results indicate the existence of optimal insider ownership, and any deviation from it causes FF to decrease. In addition, FF is higher when CEO ownership is lower (<0.08%) and this phenomenon is more pronounced for financially unconstrained firms. Furthermore, FF is higher when non-CEO insider ownership is in the middle range (0.12–0.43%) for financially constrained firms, whereas non-CEO insider ownership has minimal impact on FF for unconstrained firms.

Keywords:

• insider ownership
• financial flexibility
• corporate governance
• ownership structure

JEL Classification:

• G30
• G32

Acknowledgement

Referees’ comments are greatly appreciated.

Notes

¹ Financial flexibility is proxied by the cash flow sensitivity of external financing in this study. Schoubben and Van Hulle (2011) follow the similar approach.

² Concentrated ownership is associated with inflated earnings, whereas dispersed ownership with extraction of private benefits of control (Coffee, 2005).

³ McConnell and Servaes (1990) find a non-linear relationship between Tobin’s Q and insider ownership. The relationship is positive when the insider ownership is lower. Tobin’s Q peaks when insider ownership is in the range of 40% to 50%, after which such a relationship turns negative.

⁴ The sample of Harford et al. (2008) spans the period of 1993 to 2004 as opposed to 1992 to 2009 in our sample.

⁵ Roodman (2009) shows that the two-step robust estimator is better than the one-step counterpart.