INEGI Resources

Are you working with Mexico’s Economic Censuses?

Test and debug your code before visiting the Microdata Laboratory or requesting remote processing from INEGI.

This package generates synthetic, panel-compatible Economic Census datasets for 2004, 2009, 2014, 2019, and 2024. It is designed to help researchers develop complete empirical workflows locally before running them on confidential microdata.

Main features

• Five Census waves: Produces datasets for 2004, 2009, 2014, 2019 and 2024.

• Panel-compatible structure: Simulated datasets contain the panel identifier "id_uelm" and the IADB identifiers.  

• Real classifications: Uses valid six-digit industry codes and real geographical codes for states, municipalities, localities and AGEBs.

• Calibrated scale: Establishment counts and major economic aggregates are consistent with public data (Sistema Automatizado de Información Censal) at the municipality and broad-sector level.

• Economic coherence: Employment, capital, wages, inputs, production and value added are generated through a correlated economic structure (a noisy Cobb–Douglas production function).

• Exact accounting identities: Related variables satisfy their accounting definitions in the questionnaires and methodology, allowing code based on totals and components to be tested reliably.

• Realistic firm structure: Includes single- and multi-establishment firms.

• Server-aware datasets: The primary outputs follow the wave-specific variable inventories currently available in INEGI’s confidential environment.

• Flexible size: The default development mode simulates approximately 10% of the Census. Users can choose another percentage or run a more resource-intensive full-size simulation.

• Easy to use: One master Stata dofile, one clearly marked settings section, and no user-written Stata packages.

How to use it

Download and unzip the package, open simulate_all.do, and edit the clearly marked USER SETTINGS section. There you can select:

• The package directory

• Development or full mode

• The percentage of the Census to simulate

• Whether to save the optional expanded datasets

The primary outputs are named Insumo2004.dta, Insumo2009.dta, Insumo2014.dta, Insumo2019.dta, and Insumo2024.dta.

Requirements: Stata 16 or newer. No additional packages are required.

This package contains no confidential microdata. Its outputs are intended exclusively for developing and testing code, not as empirical evidence about Mexico.