Empirical Essays on Efficiency, Productivity, Structural Changes and Resource Use in Hawaii Agriculture

Abstract

This dissertation is a collection of empirical essays that investigate Hawaiʻi agriculture. Chapter 1 presents a general introduction and an overview of the dissertation. Chapter 2 gives an account of the present status of Hawaiʻi agriculture and, discusses the availability of agriculture data. Chapter 3 examines efficiency and productivity of vegetable farms in Hawaiʻi using data from five census years (1982, 1987, 1992, 1997 and 2007) and employing Data Envelopment Analysis (DEA). The DEA efficiency scores suggest considerable inefficiencies in Hawaiʻi’s vegetable farms. The technical efficiency of an average vegetable farm has deteriorated over the years. While the technical efficiency of small farms has decreased, the technical efficiency of large farms has increased. However, farms are found to be relatively scale-efficient. The deterioration of technical efficiency is mainly due to a decrease in pure efficiency. Malmquist Index (MI) productivity analysis reveals that aggregate productivity of Hawaiʻi’s vegetable farming sector has decreased on average. The MI decomposition shows that technical efficiency on average has declined significantly.. Chapter 4 examines productivity changes in Hawaiʻi’s vegetable sector arising from the entry, exit and survival dynamics. Melitz and Polanec (2014) extension of the dynamic Olley-Pakes (1996) productivity decomposition is employed in this analysis. The study finds that exiting and new farms show low productivities compared to continuing farms. Therefore, while farm exits contribute positively to the aggregate productivity growth, new farm entry contributes negatively to the aggregate productivity growth. The study also finds that the aggregate productivity of continuing firms has increased between census years. Productivity gains in continuing farms are found to be mainly a result of market share reallocation between farms. Chapter 5 presents a mathematical model that aids in simulating policy or environment changes under limited data conditions. Hawaiʻi crop sectors were modeled and calibrated to the base year 2007 data using the Positive Mathematical Programming (PMP) method. The study simulates the scenario of a loss of farmland to assess its effect on acreage allocation among crops. The study identifies the possibility of modeling Hawaiʻi’s agriculture for policy simulations and suggests ways to improve a state-wide agricultural model to closely represent real situations.