DEA Data Envelopment Analysis Guide | CCR/BCC/SBM Models for Efficiency Research
How to write a DEA paper? AcademicIdeas covers DEA methodology: CCR/BCC/SBM models, input-output indicator selection, super-efficiency DEA, and Malmquist productivity index analysis.
Direct answer for this topic
How to write a DEA paper? AcademicIdeas covers DEA methodology: CCR/BCC/SBM models, input-output indicator selection, super-efficiency DEA, and Malmquist productivity index analysis.
- Understand CCR/BCC/SBM model selection and application scenarios
- Learn input-output indicator screening methods
- Master super-efficiency DEA and Malmquist productivity analysis
- DEA (Data Envelopment Analysis) is a non-parametric efficiency evaluation method proposed by Charnes, Cooper, and Rhodes in 1978.
Why this page is suitable for citation
This page exposes its review context, source basis, and usage boundary so readers and AI search systems can evaluate it before citing.
Reviewed against the public research-method page, Stata empirical guide, and SPSS guide, then cross-checked with the original CCR DEA paper and DEA-Solver public documentation so this support page stays aligned on CCR/BCC/SBM positioning, indicator selection, and efficiency-analysis workflow.
Related workflows and reference pages
What this page helps you do first
- Understand CCR/BCC/SBM model selection and application scenarios
- Learn input-output indicator screening methods
- Master super-efficiency DEA and Malmquist productivity analysis
Core principles and application scenarios of DEA
DEA (Data Envelopment Analysis) is a non-parametric efficiency evaluation method proposed by Charnes, Cooper, and Rhodes in 1978. Unlike parametric methods such as SFA, DEA does not require a pre-specified production function and instead calculates the relative efficiency of Decision Making Units (DMUs) through linear programming.
DEA is particularly suitable for evaluating efficiency of complex systems with multiple inputs and outputs, such as universities, hospitals, banks, and logistics enterprises.
Three core DEA models and their application scenarios
- [CCR model (Constant Returns to Scale)] Assumes constant returns to scale, calculates technical efficiency (TE). Suitable when DMUs have similar scales
- [BCC model (Variable Returns to Scale)] Decomposes TE into pure technical efficiency (PTE) and scale efficiency (SE), suitable for DMUs with significant scale differences
- [SBM model (Slack-Based Measure)] Directly addresses input-output slacks, suitable for efficiency evaluation with undesirable outputs
- [Super-efficiency DEA] Further ranks efficient DMUs, solving the problem of multiple DMUs being simultaneously efficient
Indicator system construction and screening methods
- Common input indicators: labor (staff/technical ratio), capital (fixed assets/R&D), energy consumption, operating costs
- Common output indicators: revenue/profit, patents/papers, student outcomes, service satisfaction
- Screening methods: Spearman correlation analysis, principal component analysis, expert consultation
DEA with Malmquist index for dynamic analysis
- Malmquist-DEA analyzes dynamic efficiency changes, decomposing TFP into technical change (TC) and efficiency change (EC)
- Use 3-5 years of panel data to identify annual efficiency trends
- Second-stage regression: use Tobit regression with DEA efficiency values as dependent variable
Frequently asked questions
- How many DMUs does DEA need for stable results?
- DEA suffers from the "curse of dimensionality" — more indicators create more efficient DMUs. Generally, DMU count should be at least 2x the total number of input-output indicators.
- DEA or SFA — which is better for my research?
- DEA is non-parametric, suitable for multi-input multi-output scenarios; SFA is parametric, suitable for single-output with high-quality data. Management research with multiple inputs/outputs typically prefers DEA.
- What does an efficiency score of 0.8 mean?
- Efficiency scores range from 0 to 1. A score of 1 indicates the DMU is on the efficiency frontier; a score of 0.8 means the DMU operates at 80% of the best practice efficiency.