Technical efficiency of crop production in India: a meta-regression analysis with regional disaggregation and empirical synthesis of efficiency estimates
Sudha Kumari, Rakesh SinghPurpose
This paper is based on a meta-regression analysis (MRA) of technical efficiency (TE) in crop production across India, synthesizing 448 observations from 140 studies published between 1991 and 2024. The analysis explores the extent to which methodological, crop-specific and regional factors explain variation in reported mean technical efficiency (MTE) estimates.
Design/methodology/approach
The current research uses an MRA framework to synthesize quantitative TE estimates for crop production across India. Three estimation approaches, Tobit, fractional regression model (FRM) and restricted maximum likelihood (REML), are employed to quantify the influence of study-level characteristics.
Findings
The results show an average MTE of 0.75, indicating significant potential to improve crop production efficiency. Studies using panel data, parametric models and more input variables tend to report higher efficiency scores than those using cross-sectional data and fewer input variables. At the same time, models with multiple outputs report a lower efficiency score than single-output models. Regional disparities are substantial: the southern region reports the highest efficiency levels, whereas eastern, central, western and northeastern regions show consistently lower scores. Crop-type effects are also evident – cash crops exhibit the highest efficiency relative to oilseeds, followed by cereal and pulse crops.
Research limitations/implications
The study relies on published literature, which may be subject to publication bias and regional data gaps. Additionally, differences in study design and variable definitions across source studies may affect comparability and the precision of meta-regression estimates.
Originality/value
This study provides the first large-scale MRA of TE in Indian crop production. It contributes original insights by revealing that methodological choices – such as the use of panel data, parametric models, and the number of input variables – significantly affect reported efficiency scores, while also highlighting regional and crop-specific disparities that inform targeted agricultural policy.