This thesis consists of three chapters. In Chapter 1, we study whether access to complaintresolution systems can resolve hold-up problems in the implementation of public good projects. We run a field experiment involving 1629 low-caste local representatives who were unable to start public goods projects in their constituencies due to bureaucratic hurdles. We randomize offers to file complaints regarding public good project initiation on their behalf and track its effects. Our treat- ment leads to a 40 percentage points jump in complaint filing rate and is effective in improving project implementation: treated constituencies see a 26% rise in public good projects. We also find that the treatment increases project initiation in neighboring constituencies by 23%. Our analysis suggests that the mere threat of a formal com- plaint technology could cause project initiation in neighboring wards. However, when multiple complaints are filed against the same higher bureaucrat, resolution rates go down. Surprisingly, treated representatives do not gain any electoral returns in the local elections that were held two years after the treatment. In Chapter 2, we study the distributional consequences of reservation policies in thecontext of mandated political representation (“reservation") in favour of the marginalized Scheduled Caste (SC) groups in India. We bring to bear a wealth of data: secondary data on public goods from across 45,000 villages, private assets from over 19 million rural households, political candidacy data of over 300,000 candidates and a primary survey of over 8,700 households from the state of Bihar. Using a regression discontinuity design framework, we show that reservation for SCs for the post of local government head (a) lowers SC-non- SC disparities in access to public goods in the short-run (5 years later) and long-run (13 years later) (b) lowers inter-group private asset inequality modestly in the short-run and substantially in the long-run (c) creates different sets of winners and losers within SCs and non-SCs (d) has no efficiency consequences in the short-run and (e) increases political participation and presence of SCs in local government in the long-run. We exploit a unique feature of our RD design to show that the causal impact of reservation is largest when SCs are neither too large nor too small in number. Turning to mechanisms, we show that (i) government schemes are better targeted towards SCs in reserved constituencies and (ii) intra-SC heterogeneity lowers impacts of reservation. Chapter 3 studies why minorities are underrepresented in enterprise ownership and leadershippositions in big firms. This chapter empirically investigates the role of one potential reason for this: discrimination against minority employers by subordinate workers. I embed a field experiment in the recruitment of entry-level workers by a set of firms based in India. The field experiment aims to answer two main research questions: 1) Do minority employers face discrimination from below in labor markets? 2) What are the underlying motivations? I specifically test for two potential motives: attention discrimination and social image concerns. Preliminary results show that applicants are 3 p.p. (26%) less likely to apply for jobs advertised by minority employers. I also find strong evidence for ‘attention discrimination’ against minority employers.