Subsidizing Startups under Imperfect Information

How is firm creation affected by imperfect information? Our model of rational inattention generates inefficient entry and labor misallocation, because startups make error-prone decisions. Learning decisions prospective entrepreneurs alter the effects lump-sum transfers to startups: total employment gain amplified due an unintended increase in entry, most entrants hire fewer workers, misallocation goes up. The transfer changes learning incentives makes even productive lean towards more conservative hiring. We show that this novel information channel works against well-known mechanisms, thus can help reconcile recent empirical evidence on startup policies.