|dc.description.abstract||This dissertation is composed of three essays. The first essay examines the effect of bank branch expansion on High Yielding Variety (HYV) seed adoption by Indian households using the Additional Rural Incomes Survey (ARIS) and Rural Economic and Demographic Survey (REDS) Indian household panel data, and bank branch data from the Reserve Bank of India. I use the Indian government's social banking policy to provide exogenous variation in district bank access. This policy, in effect 1977 to 1990, forced banks to open more branches in financially less developed areas. I find that districts with lower initial financial development experienced significantly more branch openings during the social banking period (1977-1990). Moreover, I find that households in financially less developed districts were more likely to adopt HYV seeds during the social banking period consistent with the view that access to credit is a major determinant of new technology adoption.
In the second essay, following the empirical strategy developed in the first essay, I examine the effect of bank branch expansion on High Yielding Variety (HYV) seed adoption by Indian districts. I use the Indian government's social banking policy to provide exogenous variation in district bank access. This policy, in effect 1977 to 1990, forced banks to open more branches in financially less developed areas. Using data on district HYV use from the Evenson and McKinsey India Agriculture and Climate dataset, and bank branch data from the Reserve bank of India, I find that financially less developed districts increased HYV seed adoption during the social banking period, consistent with the view that access to credit is a major determinant of new technology adoption.
In the third essay (co-authored with Steven Craig, Wided Hemissi and Bent Sorensen), we study how state governments manage the finances of their Unemployment Insurance (UI) programs. The operation of the UI programs is separate from states' general budgets with clearly specified rules of saving (in a trust fund operated by the treasury) and spending, although spending includes a large discretionary component. Using a panel of US states we find that UI program spending and taxes are not described well by the PIH or by the Barro tax smoothing model. Instead, we find that states increase spending when their trust fund balance is high, and that trust fund balances seem to be clearly mean reverting. This pattern suggests that the data may be explained by a buffer stock model with forward looking but impatient politicians as suggested by Carroll (1997) for consumers. We split UI benefits spending into a compulsory part (explained by unemployment) and a discretionary part. Considering taxes as income and discretionary benefits as consumption, we calibrate and simulate a version of Carroll's buffer stock model. We find that the simulation results of the buffer stock model match well our data, where politicians adjust policy to stay close to the target level of savings as shown by simulation matching the covariance condition where consumption and savings move with differences from the target level of savings.||