Harvest Cost and Citrus Mechanization: Real Options Approach with Least Squares Monte Carlo Simulation

Abstract

The prospect of immigration policy reform has renewed growers' concerns of serious labor shortages and cost increases because a large portion (at least 53%) of the workforce in agriculture is unauthorized for U.S. employment. This concern of labor cost increase is more serious for specialty crop agriculture, not only because it is highly labor intensive, but also it requires labor in a very short period, particularly at harvest time. Agricultural employers may address the increased labor cost, especially harvest cost in various ways, but likely options include adoption of mechanical harvesting. If we focus on adoption of mechanical harvesting, the most imminent effect of immigration policy reform might be on crops mainly used for processing for which labor-intensive technology is currently employed, but an alternative, less labor-intensive technology is already developed. Florida citrus is a current example of a major specialty crop with these characteristics. In addition, the Florida citrus industry is facing many difficulties, from recent hurricanes to new diseases to increased international competition, all of which may urge immediate mechanization of harvesting operation intended for cost reduction. Currently, the estimated cost of mechanical harvesting of Florida oranges for juice processing suggests a significant cost advantage over hand harvesting, but the adoption of mechanical harvesting systems remains relatively low at about 7.5% of the Florida orange acreage. In order to study the mechanization-investment decision by the Florida citrus farmers, we firstly estimate the value for two operational modes (hand and mechanical harvesting) using the net present value (NPV) approach. While, the NPV approach simply compares NPVs from the two operational modes, the real options approach (ROA), which applies financial option theory for investment in real assets, assumes that the farmer does not invest until the NPV of the mechanical harvesting operation is greater than that of the current operation by margin of the option value of investment. Option valuation with early exercise features with multiple stochastic factors has the so-called dimensionality problem. The ROA study for the Florida citrus mechanization is one of these cases since there are at least four stochastic factors: price, yield, production cost, and harvest cost. We overcome this dimensionality problem by using the least squares Monte Carlo simulation (Longstaff and Schwartz 2001) which solves the dynamic optimization problem posed by the Bellman equation, where the value of continuation is computed on the basis of a regression on cross-sectional data.