Reputation and Disclosure in Dynamic Networks

We develop a continuous-time model of reputational disclosure in directed networks of biased intermediaries with career concerns. A payoff-relevant fundamental follows a diffusion and a decision maker chooses actions to track it. Experts obtain verifiable signals that reach the decision maker only if relayed by intermediaries. Intermediaries choose whether to forward evidence and an observable disclosure clock that controls the arrival rate of disclosure opportunities. Because clocks are public, silence is state dependent: when the clock is on, delay is informative and reputationally costly; when it is off, silence is mechanically uninformative. Disclosure becomes a real option on reputational capital. Along any expert-decision maker path, Markov perfect Bayesian equilibria are ladder policies with finitely many posterior cutoffs, and clock-off windows eliminate knife-edge mixing. With sufficiently high reputational stakes and low discounting, dynamic incentives rule out persistent suppression and guarantee eventual transmission of all verifiable evidence along the path, even when bias reversals block static unraveling. We then study network design and formation. Absent the high-reputation regime, among trees exactly the bias-monotone ones sustain disclosure. Under homogeneous reputational intensities the bias-ordered line is dynamically optimal; with heterogeneous intensities, optimal design screens by topology, placing high-reputation intermediaries on direct parallel routes rather than in series. In an endogenous link-formation game, pairwise stable networks can be inefficiently sparse or redundantly dense because agents ignore the option-value externalities their links create or destroy for others' reputational assets.