We study a long-term principal-agent problem involving experimentation with hidden actions. The principal delegates a costly experiment on the quality of the project to the agent. Since the agent is financially constrained in exerting effort, they require monetary transfers before exerting effort each time. The agent can privately divert these funds for personal use. When news arrives, the agent has the option to conceal the news arrival to continue receiving monetary transfers without exerting effort. The principal designs a time-dependent bonus to incentivize the agent to report the news. We show that a contract encouraging immediate reporting is more profitable than one that allows for delayed reporting.

I study a new sender-receiver game. The sender has two kinds of private information. One is the payoff-relevant state. The other is evidence. In my setting, the sender can communicate by revealing evidence.  The sender can manipulate evidence downwards by not upwards.  I characterize the cutoff equilibrium. Qualitatively there are three cases, depending on whether the bias is small, medium, or large. If the bias is small, only pooling equilibrium is available. If the bias is medium, there exists the cutoff equilibrium where the low-state sender conceals the evidence. If the bias is high, the sender discloses the realized evidence.   In Crawford-Sobel, a smaller absolute value of bias leads to finer information.  In my model,  a higher value of bias leads to full disclosure of the success since the sender cannot overreport. Then the role of the bias is opposite.

I consider a Bayesian persuasion problem with hidden action. When the principal can decide both information structure and monetary transfer, which relies on the stochastic output, then full separation or full pooling contract is optimal.

I investigate the fragmentation under high-frequency trading (HFT) popularity and develop a security exchange market competition model.  With friction like Bertrand competition with differentiated goods, the effect of the number of venues on the bid-ask spread depends on the business stealing effect. Using the Nikkei 225 future data and focusing on the system down of the Singapore Exchange as a natural experiment, I show that a smaller number of venues means a large bid-ask spread. As regards the welfare loss from HFT investment as well as the effect of the bid-ask spread effect, the relationship between the number of venues and welfare is vague.