Prev:
W2, Next:
W4
# Mechanism Design with Transfer
📗 General mechanism design problem with allocation \(x\left(t\right)\):
➩ Objective: \(\displaystyle\max_{x, p} f\left(x, p\right)\)
➩ Incentive Compatibility: \(R\left(x\left(t\right); t\right) - p\left(t\right) \geq R\left(x\left(\tau\right); t\right) - p\left(\tau\right)\) for every \(t, \tau \in T\).
➩ Individual Rationality: \(R\left(x\left(t\right); t\right) - p\left(t\right) \geq 0\) for every \(t \in T\).
# Single Crossing Condition
📗 Also called Spence Mirrlees property in mechanism design: \(\dfrac{\partial R}{\partial x \partial t} \geq 0\).
➩ Replace IC by FOC + SOC.
➩ Differentiate FOC and substitute into SOC to get \(\dfrac{\partial R}{\partial x \partial t} x'\left(t\right) \geq 0\).
➩ Assuming monotone allocation \(x'\left(t\right) \geq 0\), we have SCC.
➩ Given SCC, implementability iff \(x'\left(t\right) \geq 0\): differentiate \(R\left(x\left(\tau\right); t\right) - p\left(\tau\right)\) with respect to \(\tau\) and substitute in FOC of IC.
# Virtual Surplus Optimization
📗 Revenue optimization \(\displaystyle\max_{x, p} E_{T}\left[p\left(t\right)\right]\).
➩ Differentiate \(R\left(x\left(t\right); t\right) - p\left(t\right)\) with respect to \(t\), assume the derivative is larger than 0.
➩ Use FTC to write \(R\left(x\left(t\right); t\right) - p\left(t\right)\) as an integral to compute \(p\left(t\right)\).
➩ Integrate \(p\left(t\right)\) over \(t\) and do integration by parts to get \(E\left[R\left(x; t\right) - \dfrac{\partial R\left(x; t\right)}{\partial t} \dfrac{1 - F\left(t\right)}{f\left(t\right)}\right]\).
➩ When \(R\left(x; t\right) = x t\), the virtual surplus for auctions \((t - \dfrac{1 - F\left(t\right)}{f\left(t\right)})\).
Last Updated: November 25, 2025 at 1:46 AM