This dissertation examines how firm-specific behavior concerning factors of production is shaped by institutional constraints in development countries. The initial two chapters analyze how firms in Brazil compensate workers for their labor: the first centers on the role of the collective bargaining framework, and the second quantifies the impact of firms on the racial wage gap. The final chapter focuses on firms' use of credit for working capital in response to disruptive periods of violence during Mexico's Drug War. Firms compensate workers not only with wages, but also with other job characteristics that the labor literature broadly refers to as amenities. However, it is hard to study amenity compensation because we rarely observe variation in amenities across establishments in some systematic way. One exception is the comprehensive set of amenities codified in the text of collective bargaining agreements (CBAs) that unions negotiate with employers. In chapter 1, I leverage a reform that automatically extended all existing CBAs in Brazil to analyze the impact of this new collective bargaining framework on firm compensation, as measured by wages and amenities, as well as subsequent selection effects in the workforce. To quantify the value workers place on amenities secured by unions, I measure how textual elements in CBAs influence an establishment's ability to poach workers from other employers, conditional on wages, using data on the universe of CBAs merged with an administrative linked employer-employee dataset. I find that automatic extensions increase compensation by 1.6-3.8% when unions are strong---an effect that is driven by additional amenities whose value more than offsets foregone wage gains. These changes in compensation lead to an increase in hiring concentrated among low-skill workers, implying an elasticity of labor supply to the affected firms of around 2. Further evidence suggest that unions reduce compensation inequality within establishments. While union-driven changes to firm compensation can lead to an influx of low skill workers, how firms select and pay workers can have important consequences for wage disparities between groups. In Chapter 2 (work co-authored with François Gerard, Edson Severnini, and David Card), we measure the effects of firms' employment and wage setting policies on racial pay differences in Brazil. We find that nonwhites are less likely to work at firms that pay more to all race groups. This sorting pattern explains about 20% of the white-nonwhite wage gap for both genders. Moreover, the pay premiums offered by different employers are also compressed for nonwhites relative to whites. This within-firm differential wage setting contributes another 5% of the overall gap. We then explore to what extent the under-representation of nonwhites at higher-paying firms is due to the selective skill mix at these workplaces. Using a counterfactual based on the observed skill distribution at each firm and the nonwhite shares in different skill groups in the local labor market, we conclude that assortative matching accounts for about two- thirds of the underrepresentation gap for both men and women. The remainder reflects an unexplained preference for white workers at higher-paying firms. Interestingly, the wage losses associated with unexplained sorting and differential wage setting are largest for nonwhites with the highest levels of general skills. This suggests that the allocative costs of race-based preferences may be relatively large in Brazil. The first two chapters reveal that firms exercise some discretion over compensation and hiring within the context of institutions such as collective bargaining and nondiscrimination laws. But firms are also constrained by other institutions in how they carry out their day-to-day activities. In particular, the capacity of the State to exercise control over the legitimate use of force promotes the fundamental trust required between agents to make welfare-enhancing transactions. In Chapter 3, I analyze how drug-related violence affects credit use by micro and small enterprises (MSEs). Leveraging administrative data on working capital credit lines issued to MSEs in Mexico, I exploit geographic variation in homicide rates as well as exogenous kingpin captures to identify the causal effects of violence on credit use. I find that firms significantly increase the amounts drawn from their credit lines after experiencing violence shocks. More credit use could be motivated by rising short-term liquidity needs (distress story) or increasing risk of holding cash (substitution story). Rising default probabilities indicate signs of distress, although heterogeneity analyses reveal cash-for-credit substitution among non-revolving borrowers. I also find evidence that rising liquidity needs among distressed MSEs are likely driven by decreased economic activity rather than theft or extortion. As such, this paper highlights the important role that financial products play in terms of helping firms absorb violence shocks as well as providing safe alternatives to cash holdings under insecure environments.