Public agricultural research and research spillover effects on the 48 United States states' agricultural productivity: Cluster, period and overall analyses

This dissertation attempts to apply a clustering procedure to group 48 U.S. states (excluding Hawaii and Alaska) into clusters, to compute productivity indexes for intertemporal, spatial and temporal-spatial comparisons and to identify the sources of productivity variations across states and/or over time.; The period of 1960--1990 was divided into 5 sub-periods. The 48 U.S. states were grouped into 5 clusters for each sub-period based on the similarities in their characteristics of agricultural production. Three types of agricultural productivity were computed: Fisher intertemporal, spatial and temporal-spatial productivity indexes. Mixed-effect models were used to relate public transportation, public and private agricultural research, research spillovers, public extension, government farm programs, off-farm wages, labor quality, milk price support programs, farm sizes, temperatures and rainfall to a state's intertemporal, spatial and temporal-spatial agricultural productivity, respectively. The effects of these factors were evaluated at the overall, cluster and subperiod levels and were assumed homogeneous across states except for public agricultural research, research spillovers and government farm programs.; Results showed the identification of five clusters: poultry/cotton, vegetable/fruit, livestock/crop, livestock and dairy clusters. For states that were clustered together, agricultural productivity did not necessarily grow at the same speed nor were the states equally productive.; Public and private agricultural research was always positive and significant at the overall and cluster levels. Research spillovers, from states in the same cluster were always significant and positive in the dairy cluster. Research spillovers from states in different clusters were never significant in the livestock/crop cluster. Both total research spillovers and spillovers from different clusters were always significant in the periods of 1966--1971, 1972--1977 and 1978--1983. The density of interstate highway network was positive and significant at the overall level but the effect appeared to vanish at the cluster or sub-period level. Public extension per dollar of farm sales always had a significant and a negative effect. Off-farm wages were always significant and negative at the overall level and in the last two sub-periods. Government farm programs were always significant and positive at the overall level. Labor quality was always significant and positive at the overall level and in the livestock/crop cluster. NEW price support was never significant in the livestock/crop cluster but was always significant and negative during 1972--1977.