We appreciate comments from workshop participants at Duke University, London Business School, Seoul National University, University of Technology, Sydney, and Stockholm School of Economics, and from Campbell Harvey, Will Mitchell, and Mohan Venkatachalam. Financial support was provided by the Fuqua School of Business and the Global Capital Markets Center at the Fuqua School of Business. Analyst forecast data are from Value Line. We are grateful for the excellent research assistance of Mark Evans and Kevin Ow Yong.
Voluntary Disclosure, Earnings Quality, and Cost of Capital
Article first published online: 2 JAN 2008
Journal of Accounting Research
Volume 46, Issue 1, pages 53–99, March 2008
How to Cite
FRANCIS, J., NANDA, D. and OLSSON, P. (2008), Voluntary Disclosure, Earnings Quality, and Cost of Capital. Journal of Accounting Research, 46: 53–99. doi: 10.1111/j.1475-679X.2008.00267.x
- Issue published online: 2 JAN 2008
- Article first published online: 2 JAN 2008
- Received 22 September 2004; accepted 8 July 2007
We investigate the relations among voluntary disclosure, earnings quality, and cost of capital. We find that firms with good earnings quality have more expansive voluntary disclosures (as proxied by a self-constructed index of coded items found in 677 firms' annual reports and 10-K filings in fiscal 2001) than firms with poor earnings quality. In unconditional tests, we find that more voluntary disclosure is associated with a lower cost of capital. However, consistent with the complementary association between disclosure and earnings quality, we find that the disclosure effect on cost of capital is substantially reduced or disappears completely (depending on the cost of capital proxy) once we condition on earnings quality. Extensions probing alternative proxies show that our findings are robust to measures of earnings quality and cost of capital, but not to other measures of voluntary disclosure. In particular, we find opposite relations for voluntary disclosure measures based on management forecasts and conference calls, and we find no relations for a press release based measure.