We study whether investor voice affects firm policy. Using a novel dataset of the transcripts of online earnings communication conferences (OECC) of Chinese listed firms during 2006 and 2015, we show that firms do respond to investor complaints about dividend policy with increased future dividends. We find that the number of dividend‐related questions raised by investors, the number of complaints (questions with negative attitudes), firm executive responsiveness, and CEO attendance in the OECC are positively associated with future dividend payouts. Moreover, positive responses are concentrated in firms facing greater pressure from peers, the market, and the regulator. Our additional analysis shows that the increase in payouts in response to investors' voice is not at the expense of cutting value‐increasing investment. Overall, our results suggest that the online interactive platform empowers shareholders' voice and helps them to conduct a successful dialogue with firm managers in a cost‐effective way.

This study investigates the effect of decentralisation of SOEs on stock price crash risk. In so doing, we test two competing hypotheses. Under the Political Influence Hypothesis, decentralisation aggravates local government's expropriation of minority shareholders (type II agency conflict), and thus increases crash risk. Under the Local Information Hypothesis, decentralisation decreases monitoring distance (type I agency conflict), strengthens external monitoring and thus decreases crash risk. We find robust evidence supporting the Political Influence Hypothesis. Cross‐sectional analyses show that our baseline results are more pronounced when firms are decentralised to the provincial level and politicians have greater incentives to pursue their political objectives. We further show that bad news hoarding and risk‐taking are two potential channels through which SOE decentralisation increases crash risk. Taken together, our results imply that the decentralisation exacerbates the type II agency conflict rather than ameliorates the type I agency conflict in SOEs.