Impact of irrigation and fertilizers on the growth of output in Andhra Pradesh: A Bayesian approach

Concerns estimation of 'production functions' for regions (e.g., a State) when time-series data on outputs and inputs are available for sub-regions (e.g., districts). An approach is proposed and applied which is an alternative to working with aggregate regional data and which derives regional functions from estimates for sub-regions. Not only is the resulting relationship meaningful as a description of the regional input-output relationship but it also has certain attractive statistical properties based on optimal extraction of information from the sub-regional functions. Proceeding on the assumption that statistical macro production functions have some use, a Cobb-Douglas type specification for four major crops (rice, sugarcane, groundnuts and tobacco) of the State of Andhra Pradesh in India is estimated. The available data restrict the analysis to only four inputs: irrigated and unirrigated land, fertilizers and rainfall. -from Authors