Assignment mechanisms: Common preferences and information acquisition

I study costly information acquisition in a two-sided matching problem, such as applicants to schools. An applicant's utility is sum of common and idiosyncratic components. The component unknown the applicant but can be learned at cost. As learn, their preferences over schools become more heterogeneous, improving match quality. In my stylized environment, too few acquire an ordinal strategy-proof mechanism. Subsidies, disclosure applicants' priorities, affirmative action-like policies lead higher Pareto improvements. Learning may also decrease when mechanism replaces Immediate Acceptance