Stochastic Frontier Analysis of the Romanian Food Industry

Abstract

This study aims to investigate the Romanian food industry companies' performance. The performance analysis was performed using stochastic frontier analysis (SFA), which was based on the financial statements of 1 464 Romanian food industry firms between 2018 and 2020. The companies operate in different sub-sectors of the food industry. In contrast to DEA, stochastic frontier analysis (SFA) requires a production function to evaluate efficiency. Gross value added was used as the result variable in the production function. The explanatory variables of the efficiency model were material costs, employee expenses, depreciation and amortisation, and other operating expenses. The results of the SFA calculation show that all regression coefficients are significant at less than 0.01% level. Based on the SFA analysis results, it can be concluded that 87% of the examined food companies have achieved at least 50% efficiency in all three years. The average efficiency scores of all subsectors are above 0.66, which can be considered good enough. Then some classification criteria (as 'z' variables) were included in the SFA model considering heterogeneity, such as region, sub-sector, and size based on the number of employees. Considering these factors, the average efficiency of food industry companies improve significantly by 26.51% in all three years. SFA analysis by sectors shows that companies operating in fruit and vegetable processing have the lowest average efficiency (0.60). The highest efficiency can be observed in other food products (0.71) and milk and dairy processing (0.69) companies. Enterprises with more than 250 employees (0.73) and 1-4 employees (0.71) are more efficient than medium-sized enterprises. The efficiency scores also show moderate differences between regions with the lowest and highest efficiency (9.37%).