VXe๎๑n@ะ๏Hwๆ
@@ฝฌQWNx@@@@t@JeBEZ~i[
@@@@@@@@@@@@@@@iFaculty Seminarj
๚@@@@F@ฝฌQXNPPV๚iฮj
PUFRO`PWFOO
๊@@@@F@RFPQOO|L
i๏า@@@F@I์@ท๕
u@@า@F@Georgy
Artemov@i{ๅwoฯwj
@่@@@F@Assignment
mechanisms: common preferences and information acquisition
@@@@@@@@@@@@@@@@@@
abstract
In this
paper I study costly information acquisition in a two-sided matching problem,
such as the assignment of students to schools. Each studentfs utility from
attending a school is a sum of a common and a student-specific components. The
common component is known by all students. The student-specific component is
unknown to a student, but can be learned at a cost. When an ordinal
strategy-proof mechanism, such as Deferred Acceptance (DA), is used, fewer
students than socially optimal acquire information (under general conditions).
Moreover, the welfare gain from introducing the DA in place of the random
assignment may be smaller than the gain from augmenting the DA by policies that
achieve the socially optimal level of information acquisition. The DA is often
promoted as a replacement of the popular Immediate Acceptance (IA) mechanism.
However, because the IA often provides higher incentives to acquire information,
the welfare gain from replacing the IA by the DA would be routinely overstated,
unless DA is augmented by the policies which promote information acquisition. I
discuss some of such policies that are used in real-life school assignments.