We study a dynamic model of voluntary disclosure of information by a potentially informed agent. The extant theoretical literature on voluntary disclosure focuses on static models in which an interested party (e.g., a manager of a firm) may privately observe a single piece of private information (e.g., Grossman 1981;Milgrom 1981;Dye 1985;and Jung and Kwon 1988) or dynamic models in which the disclosure timing does not play a role (e.g., Shin 2003Shin , 2006 as the manager's decision is what to disclose but not when to disclose it. Corporate disclosure environments, however, are characterized by multi-period and multi-dimensional flows of information from the firm to the market, where the information asymmetry between the firm and the capital market can be with respect to whether, when, and what relevant information the firm might have learned. For example, firms with ongoing R&D projects can obtain new information about the state of their projects, where the time of information arrival and its content is unobservable to the market. This is common, for example, in pharmaceutical companies that get results of a drug's clinical trial (prior to FDA approval). Such results are not required to be publicly disclosed in a timely manner and investors' beliefs about the result of a drug's clinical trial may have a great effect on the firm's price. The multidimensional nature of the disclosure game (multi-period and multisignal) plays a critical role in shaping the equilibrium; e.g., when deciding whether