Abstract:
One of the important questions in mechanism design is the
implementability of allocation rules. An allocation rule is called
implementable if for any agent, benefit from revealing its true type is
better than benefit from lying. I’ll show some illustrative examples.
Obviously, some allocation rules are not implementable. Rochet’s theorem
states that an allocation rule is implementable iff it is cyclically
monotone. During the talk, I’ll present a new convenient topological
condition that guarantees that cyclic monotonicity is equivalent to
ordinary monotonicity. The last one is easy to check (in contrary to
cyclic one). Graph theory and algebraic topology appear to be very
useful here.