Firm Responses to State Hiring Subsidies: Regression Discontinuity Evidence from a Tax Credit Formula