Essays on Agricultural and Regional Development

In a world of imbalance, food consumption exhibits great diversity among regions and countries. Although farmers in developed economies benefit from up-to-date agricultural technology and produce much more than they consume, households in the developing world are still combating food insecurity. This dissertation is composed of two manuscripts. One is about consumption in developing countries, while the other is related to promoting agricultural production in a developed economy.

Chapter 1 applies a three-stage demand system to nationally representative household survey data to identify food demand behavior with an emphasis on food staples in two West Africa countries ‒ Niger and Nigeria. The third stage of the demand system offers demand elasticities of specific staple items. Instead of treating the population as a whole, the study distinguishes rural and urban households and households of different welfare status. Results confirm the complexity of the food and staples demand between rural/urban areas and among welfare quintiles. Therefore, researchers and policymakers should consider not only the average demand response but also its distribution among households. In addition to demand elasticities, the effects of household demographic characteristics on the structure of food consumption are also obtained.

Chapter 2 estimates the rates of return to Virginia's public expenditure on agricultural research and extension (RandE) during 1949-2016 and attempts to address the ad hoc model selection problem common in previous studies. Among the econometric modeling strategies in previous literature, Bayesian Model Averaging (BMA) and Bayesian Hierarchical Model (BHM) are two promising methods to solve the issue of model uncertainty. The rate-of-return estimates by BHM are preferable because BHM imposes fewer restrictions on lag structures and offers more reasonable lag shapes. By BHM, the internal rates of return (IRR) of Virginia's public expenditures on agricultural RandE are 26% and 42%, respectively. Nineteen percent of Virginia's agricultural productivity growth during 1949-2016 results from its RandE investments, and the contribution of research to that growth is about twice of that of extension. One extra million dollar expenditure on research in 1992 would have brought a benefit of $4.5 million, and the same expenditure in 1983 would have brought $5.4 million in additional benefits. If the extra expenditure is spent on extension, it would have brought a benefit of $6.1 million and $6.3 million if the expenditure occurs in 1992 and 1983, respectively. Besides the modeling strategy, this study is distinguished from previous studies in that distributions of rates of return instead of only point estimates are obtained, which is missing in most studies.