Farm productivity is important since it determines the quantity of food supplied and the income of farms in the economy. Using two 5-year farm-level panel dataset that is rich in farm-specific characteristics, our goal is to explore the relationship between farm characteristics and farm productivity. For this goal, first, we estimate farm-specific productivity. Given estimated productivity, second, we select characteristics that substantially explain variations of farm-specific productivity across time, using Least Absolute Shrinkage and Selection Operator (LASSO), an emerging variable selection method. The results indicate that farm types, and income and expenditure variables, such as primary farming, subsistent farming, profit margin, operator’s age, and rent are most strongly associated with farm-specific productivity.