Menu ▾ ▴

Home

Choonjoo Lee

Title

DEAS v2.0: unified Stata command for Data Envelopment Analysis

Announcement

DEAS v2.0 is now available as a major update to the Stata Data Envelopment
Analysis package.

The main change is that the package is now organized around one unified command:

    dea ..., model(...)

The command keeps the original DEA-style syntax:

    dea inputs = outputs, options

and also supports an extended Stata-style syntax:

    dea outputs, inputs(inputs) options

Both forms produce the same results. This means older do-files remain usable,
while new work can use a more explicit option-based style.

Recommended first download
--------------------------

Start with:

    deas_v2_0_stata_package.zip

This file contains the ado-files, help files, quick-start guide, one teaching
dataset, and the main article example do-file. Most users only need this file
to begin using the new command.

Other downloads
---------------

    deas_v2_0_article_supplement.zip

contains the author-prepared forthcoming Stata Journal article materials, online
supplement, figures, and LaTeX source. Article manuscript files are covered by
ARTICLE_REUSE_NOTICE.md, not the software license.

    deas_v2_0_replication_examples.zip

contains the article examples, OECD education application, validation scripts,
logs, datasets, and figure-generating do-files.

    deas_v2_0_full_archive.zip

contains the complete accepted-file archive.

Quick start
-----------

After downloading and unzipping deas_v2_0_stata_package.zip, open Stata in the
unzipped folder and type:

    adopath ++ "`c(pwd)'"
    which dea
    help dea

Then run the included teaching example:

    use dea_6dmu.dta, clear
    dea i_employee i_area = o_sales o_profits, rts(crs) gen(eff_crs)
    dea i_employee i_area = o_sales o_profits, rts(vrs) gen(eff_vrs) reference
    dea i_employee i_area = o_sales o_profits, model(sbm) rts(vrs) gen(eff_sbm)

Main features
-------------


- Sixteen DEA model families through one dea command.
- Backward compatible inputs = outputs syntax.
- Extended syntax using inputs().
- CCR/BCC radial technical efficiency.
- SBM, additive model, radial super-efficiency, and Super-SBM.
- Cost, revenue, profit, and allocative efficiency.
- Directional distance function.
- FDH, imprecise DEA, fuzzy DEA, congestion, and undesirable outputs.
- Malmquist, window, network, cross-efficiency, and bootstrap extension commands.
- Fast Mata simplex engine and optional Stata LinearProgram() IPM engine.
- engine(auto) fallback for numerical robustness.
- Standardized r() results for easier post-estimation workflows.

Notes on game-theoretic commands
--------------------------------

The release also includes game-theoretic DEA commands for cross-efficiency
games, Nash-style bargaining, Stackelberg-style two-stage diagnostics, and
cooperative allocation. These are included as research-oriented diagnostic
tools. They should not be read as fully validated structural game estimators.

Documentation
-------------

Use:

    help dea

for the main command. Extension commands have separate help files, for example:

    help dea_malmquist
    help dea_crosseff
    help dea_bootstrap
    help dea_fuzzy

The forthcoming Stata Journal article and online supplement provide additional
model descriptions, implementation details, computational benchmarks,
validation notes, and OECD education examples.

Copyright and reuse
-------------------

The Stata software files are distributed under AFL-3.0; see LICENSE_CODE.txt.
Article manuscript files, if included, are author-prepared
forthcoming/accepted-manuscript materials and are governed separately by
ARTICLE_REUSE_NOTICE.md. The final published Stata Journal PDF should not be
redistributed through SourceForge or GitHub unless publisher permission or
policy explicitly allows it.

Feedback
--------

Please use the SourceForge Discussion or Support page for bug reports,
installation problems, unclear examples, or requests for additional worked
examples.

Short Announcement Variant

Use this shorter version if the Discussion page should stay compact.

DEAS v2.0 is now available.

This is a major update centered on one unified Stata command:

    dea ..., model(...)

It supports sixteen DEA model families, keeps the original inputs = outputs
syntax, adds an extended inputs() syntax, and includes both a fast Mata simplex
engine and Stata's LinearProgram() IPM engine.

New users should start with:

    deas_v2_0_stata_package.zip

It contains the ado-files, help files, quick-start guide, sample dataset, and
article example do-file.

After unzipping, open Stata in the package folder and run:

    adopath ++ "`c(pwd)'"
    help dea
    use dea_6dmu.dta, clear
    dea i_employee i_area = o_sales o_profits, rts(crs) gen(eff_crs)

Additional archives provide author-prepared forthcoming Stata Journal article
materials, the supplement, replication examples, validation scripts, logs,
datasets, and the full accepted-file archive. Article manuscript files are not
covered by the software license; see ARTICLE_REUSE_NOTICE.md.

Project Admins:

Auth0 Logo