Jadson16/PolicySpace

An agent-based model to evaluate fiscal redistribution among municipalities

PolicySpace

PolicySpace2 -- in-development -- now available at http://www.github.com/bafurtado/policyspace2

Please, cite as: FURTADO, Bernardo Alves. PolicySpace: agent-based modeling. IPEA: Brasília, 2018.

Available here in English: http://www.ipea.gov.br/portal/index.php?option=com_content&view=article&id=33132&Itemid=433

Also available in Portuguese: http://www.ipea.gov.br/portal/index.php?option=com_content&view=article&id=32743&Itemid=433

This is an open agent-based model with three markets and a tax scheme that empirically simulates 46 Brazilian metropolitan regions.

Developed by Bernardo Alves Furtado, funded by Institute of Applied Economic Research (IPEA). The author acknowledges receiving a grant of productivity by National Council of Research (CNPq). Isaque Daniel Rocha Eberhardt and Francis Tseng were fundamental contributors (see below for participation details).

This work is licensed under GNU General Public License v3.0

English summary

PolicySpace: a modeling platform. (accepted to) https://arxiv.org/abs/1801.00259 CAPS 2018: Complexity and Policy Studies UNCC - George Mason University - April 18 (but not presented).

Repository of produced texts

https://www.researchgate.net/profile/Bernardo_Furtado

Current collaborators

Bernardo Alves Furtado, Francis Tseng

Previous collaborators

Bernardo Alves Furtado, since beginning:(conception, design, coding - agents, markets, timeframe, general) Isaque Daniel Rocha Eberhardt, 16 months:(design, coding - 'controls', 'plottings', 'parameters') Alexandre Messa (some suggestions) [4 months] Davoud Taghawi-Nejad (early-on pontual suggestion)

How do I get set up?

We recommend using conda and creating an environment that includes all libraries simultaneously. Type on a terminal, after having downloaded and installed conda.

conda create -n e36 python=3.6 gdal fiona geopandas numpy mkl scikit-learn numba joblib click click-plugins cligj cycler descartes llvmlite munch numba pandas pyparsing pyproj python-dateutil pytz scipy shapely joblib Flask Flask-WTF WTForms psutil -c conda-forge

Using the above code will install the libraries. Then, you have to activate the newly created environment with the command: activate <your_env>. If any of the libraries were not available on conda-forge, use pip install -U scikit-learn, for example.

Don't forget!

Extract/Unzip file mun_ibge_2014_latlong_wgs1984_fixed.zip

How to run the model

Configuration

To locally configure the simulation's parameters, create the following files as needed:

• conf/run.py for run-specific options, e.g. OUTPUT_PATH for where sim results are saved
• conf/params.py for simulation parameters, e.g. LABOR_MARKET.

The default options are in conf/default/, refer to those for what values can be set.

Parallelization and multiple runs

These optional arguments are available for all the run commands:

• -n or --runs to specify how many times to run the simulation.
• -c or --cpus to specify number of CPUs to use when running multiple times. Use -1 for all cores (default).

Running

python main.py run

Example:

python main.py -c 2 -n 10 run

Sensitivity analysis

Runs simulation over a range of values for a specific parameter. For continuous parameters, the syntax is NAME:MIN:MAX:NUMBER_STEPS. For boolean parameters, just provide the parameter name.

Example:

python main.py sensitivity ALPHA:0:1:7

Will run the simulation once for each value ALPHA=0, ALPHA=0.17, ALPHA=0.33, ... ALPHA=1.

Example:

python main.py sensitivity WAGE_IGNORE_UNEMPLOYMENT

Will run the simulation once for each value WAGE_IGNORE_UNEMPLOYMENT=True & WAGE_IGNORE_UNEMPLOYMENT=False.

You can also set up multiple sensitivity runs at once.

For example:

python main.py sensitivity ALPHA:0:1:0.1 WAGE_IGNORE_UNEMPLOYMENT

is equivalent to running the previous two examples in sequence.

Distributions

Runs simulation over a different distribution combinations: ALTERNATIVE0: True/False, FPM_DISTRIBUTION: True/False.

Example:

python main.py -n 2 -c 2 distributions

ACPs

Runs simulation over a different ACPs.

Example:

python main.py -n 2 -c 2 acps

Regenerating plots

You can regenerate plots for a set of runs by using:

python main.py make_plots /path/to/output

In Windows, make sure to use double quotes " " and backward slashes as in:

python main.py make_plots
"..\run__2017-11-01T11_59_59.240250_bh"

Running the web interface

There is a preliminary web interface in development.

To run the server:

python main.py web

Then open localhost:5000 in your browser.

---

Last written documentation

arxiv.org/1702.03226v2.pdf AAMAS/ABMUS text arxiv.org/1609.03996v1 SEAL operational manual

Major changes since then

1. Firms no longer have FIXED salaries scheme. Really depends on sale. Major general improvements
2. Five TAXES introduced: consumption, labor, firm, real estate transaction, property
3. Also FOUR schemes of TAXES distribution among municipalities: multiple regional testing and modeling of taxes distribution (multi_ACP_alternate)
4. Endogenous prices setting in a sticker manner
5. Decision-making of the firm profits hire, otherwise fire

Next major move

TRANSPORT SYSTEMS or initial approach (such as BUS or CAR choices, related to income and time/productivity)

MAJOR misses

1. Credit market
2. Exogenous regional MIGRATION
3. Housing rent market
4. Sectors and products variety

Initial reference

• Based on Lengnick, 2013. Agent-based macroeconomics: A baseline model.
• Journal of Economic Behavior & Organization

Distinctions from Lengnick, 2013

1. Households are FIXED, but families can move into other households
2. We have either four or one government institution that collects taxes and improves quality of life index
3. Wage is paid to the agent, not on the household. However, consumption is based on families' average money
4. There is no fixed structure for the working-consumption network. Distance is used instead.
5. Different wage mechanism

What is this repository for?

• Public Policy analysis