This dissertation presents a three-part study in development economics. Each chapter investigates how technological innovations enable firms and individuals to overcome problems of asymmetric information - whether it be monitoring devices to help firms overcome problems of moral hazard; online job-boards to overcome search and matching frictions for job-seekers; or new agricultural techniques to increase information diffusion within social networks.

The first chapter examines the problem of moral hazard in employer--employee contracting and how this may be an important barrier to firm efficiency and growth in the developing world. To do so, we run an RCT with a fleet of 255 minibuses (matatus) in Nairobi, Kenya, where we introduce monitoring devices that track real-time vehicle location, daily productivity, and safety statistics. We randomize whether minibus owners have access to these monitoring data using a novel mobile app. This information allows owners in the treatment group to observe a more precise signal of driver effort, the amount of revenue drivers collected in fares, and the extent to which the driver engages in reckless driving. We find that treated vehicle owners modify the terms of the contract by decreasing the rental price they demand. Drivers respond by working more hours, decreasing behaviors that damage the vehicle, and under-reporting revenue by less. These changes improve firm profits and reduce management costs, thereby helping treated firms grow. The device also improves owners' trust in their drivers, which drivers say makes their job easier. Finally, we investigate whether these gains to the company come at the expense of passenger safety, in an environment where accidents are common. While we do not find any evidence that conditions deteriorate, offering detailed information on driving behavior also does not \emph{improve} safety. Only by incentivizing drivers through an additional cash treatment do we detect safety improvements.

The second chapter investigates how agents in a social network can be encouraged to obtain information from outside their peer groups. Using a field experiment in rural Bangladesh, we show that demonstration plots in agriculture --- a technique where the first users of a new variety cultivate it in a side-by-side comparison with an existing variety --- facilitate social learning by inducing conversations and information sharing outside of existing social networks. We compare these improvements in learning with those from seeding new technology with more central farmers in village social networks. The demonstration plots --- when cultivated by randomly selected farmers --- improve knowledge by just as much as seeding with more central farmers. Moreover, the demonstration plots only induce conversations and facilitate learning for farmers that were unconnected to entry points at baseline. Finally, we combine this diffusion experiment with an impact experiment to show that both demonstration plots and improved seeding transmit information to farmers that are less likely to benefit from the new innovation.

Finally, the third chapter explores the impact of online job platforms on labor market frictions in India. In recent years innovative job market information systems have emerged in order to match recent graduates with employers via integrated websites and call center based platforms. We use a randomized control trial to evaluate the ability of such job-portals to ease search frictions in India. We partner with Job Shikari, an online platform operating primarily in Northern India, and upload a randomly selected sample of recent vocational training graduates onto their platform. A second subset of graduates are uploaded to the platform, and receive ``priority status", which means they subsequently receive many more text messages than their peers about potentially job opportunities. We find that being uploaded to the job portal has a negative impact on the probability of being employed, but no significant impacts on job-search. The treatment priority group experiences a less strong disemployment effect, but is much more likely to respond to the treatment by migrating to urban centers. These results differ by job-seekers' observable characteristics such as age and martial status. We interpret our results as evidence that the impact of job-portals depends significantly on job-seekers beliefs about what the job-portal can do for them, rather than just on how well the platform can match the job-seeker with a particular job.

Development requires the establishment of a functioning state. To strengthen state capacity, governments in low income countries must raise tax revenue while maintaining political stability. The risk of inciting political unrest when attempting to increase taxes may trap governments in a low-tax equilibrium, but public goods provision may improve both tax compliance and political stability. To test these questions empirically, in Chapter 1, I partner with the national tax authority and a local mayor’s office in Haiti to cross-randomize both tax collection and public goods across one of the country’s largest cities. Effects are measured both via administrative data on tax revenue as well as through novel measures of political unrest. In the chapter’s main result, I show that hand-delivering property tax invoices reduces individual tax compliance by 48%, and increases independently observed measures of localized political violence by 192%. In contrast, providing a valuable and visible public good (namely municipal garbage removal) increases tax compliance by 27%, and reduces localized political violence by 85%. Importantly, public goods provision significantly mitigates the adverse effects of tax collection in neighborhoods receiving both treatments. A cost accounting exercise suggests that providing the public good in this setting could pay for itself within the first year. These findings suggest that it may be possible to peacefully shift to a new equilibrium of higher tax compliance with a sufficient initial investment in public goods provision perhaps financed through foreign aid or other transfers.

State formation also requires navigating the at times competing interests of society and the state. While social norms have been successfully leveraged to support governments in established states, different norms may be exist in settings of early state formation characterized by low rates of pro-social coordination and limited state credibility. Appeals to social norms in such settings may result in backlash as society reveals itself to be an adversary to state formation. Conversely, the government may be able to shift prevailing norms by first providing relevant services. I test these questions empirically in Chapter 2 by extending the experiment in the previous chapter with individual-level randomization over the universe of properties in my study city’s tax registry. Effects are measured by matching previously siloed administrative data. Through extensive original qualitative and quantitative data collection, I document differences in social norms and develop novel corresponding treatments. In the chapter’s main result, I show that appealing to social norms by increasing social exposure of the tax compliance decision further reduces individual tax compliance beyond the effects found in the previous chapter leading to a net loss of 66%. I find evidence that this negative effect potentially reverses when randomly combined with receiving an increase in public services. I interpret this as evidence of antisocial punishment, a mechanism from psychology that I introduce to the tax literature. I indirectly test other motivations for tax compliance by randomizing the framing of tax collection engagement. I find evidence that each appeal tested reduced the adverse tax collection response even in the absence of public goods. The largest response I find is from appealing to local proverbs that make salient obligations and shared responsibility which result in net tax payment increases of 40%. A cost accounting exercise cautions that the appeal to social exposure could have lost the city 3/4 of its property tax revenue but that the culturally relevant appeal to obligations could have resulted in $210,000 in additional annual revenue. These findings highlight the need for the state to learn when it can rely on society’s support as an ally and how to avoid society’s backlash as an adversary.

Due to the acute challenges of governance in low income countries, donors are increasingly promoting public–private partnerships to provide public goods. However, these hybrid arrangements create a different set of governance challenges as officials seek to hold private sector actors accountable for delivery. In Chapter 3, I examine one such effort in Haiti where the government attempted to achieve universal primary education by providing 200 million USD in subsidies directly to education entrepreneurs. I find evidence that the program’s roll–out coincided with a more than 30 percentage point increase in primary school enrollment and use a cohort study to identify causal evidence of both increased enrollment as well as improvements in household welfare. I make use of multiple novel datasets and a natural experiment to find that audited participants reduced grant claims by 18.9% – interpreted as a reduction in fraud – resulting in a directly observed savings of 875,000 USD per year for a conservative return of 11 USD for every dollar spent on auditing. I do not find any evidence of spillover effects. I provide additional support for my findings through an event study. These findings suggest that public-private partnerships can be a means for low- capacity states to quickly increase public goods provision, and though such an approach risks substantial fraud, relatively simple to implement interventions can cost-effectively mitigate some of those losses.