In the first chapter, we consider the problem of first-best implementation in dynamic, interdependent-value environments where agents have quasi-linear preferences and private types are correlated. Our first candidate solution (I.M.) relies on methods typically seen in games with imperfect monitoring. We derive conditions under which this approach gives the planner the flexibility to implement efficient mechanisms that are ex-post budget balanced and collusion proof. If agents are required to make one final payment after they exit the system, the I.M. mechanism can also satisfy individual rationality and adapt to finite horizon settings. When such payments cannot be enforced, but a cyclic monotonicity condition is satisfied, the planner may adopt our second mechanism (P.A.). This alternative approach employs a technique originally developed for surplus extraction in static auctions. It allows the planner to align every agent's preferences with his own and implement socially optimal outcomes while satisfying ex-ante budget balance and individual rationality. Both mechanisms permit the planner to implement any sharing rule when redistributing surplus.

In the second chapter (co-authored with Sambuddha Ghosh), we examine a paradox of strategic voting. Consider an electorate whose individual rankings of alternative policies may change between the time they vote for a candidate and the date a policy is implemented. Rankings may change following common or idiosyncratic shocks. Voters choose, via a simple majority election, between a candidate who is committed to one alternative, and an unbiased candidate who implements the ex-post optimum. We show that, even when idiosyncratic shocks lead to slight changes in rankings, the unique symmetric pure strategy Nash equilibrium often entails strategic voters choosing the committed candidate. The resulting welfare loss increases with the likelihood of a common shock.

In the third chapter (co-authored with Konstantinos Rokas), we examine an electoral framework in which agents can strategically choose when to vote and can observe the votes of those who precede them. This is in contrast to the literature on Condorcet's Jury Theorem which has limited its attention to simultaneous and sequential voting games. We show that in a common value setting, such an election aggregates information efficiently regardless of the voting rule or the size of the electorate. The proposed election format ensures that the Jury Theorem obtains for an electorate with differentially informed agents. Under more restrictive conditions, this framework can also produce the efficient outcome in environments where voters have conflicting interests or the size of the electorate is uncertain. Finally, we examine the performance of the election when private signals are multidimensional.