search

jsoma / data-studio-projects

12 stars 18 forks source link

Measuring the effects of the political reforms in Brazilian Congress [Project] #103

Closed RodrigoMenegat closed 6 years ago

RodrigoMenegat commented 7 years ago

Pitch

Summary

Representatives are discussing a political reform in the Brazilian electoral system. One of the most polemical aspects is a possible change in how the elections for the legislative power work.

Currently, seats are divided among the parties according to the percentage of the total votes each affiliation got. This means that not necessarily all the most voted representatives are elected, since what matters essentially is the votes each party receives.

There are two proposals for changing the system. The first one, called 'distrital misto', advocates that 50% of the seats should still be distributed like they are today while the other 50% should be distributed to the most voted individual candidates. The other one, nicknamed "distritão", advocates that all the seats should be given to the most voted individual candidates.

Political commentators claim that this changes may favor richer and famous candidates, making the Congress less diverse and contributing to undermine the representativeness of political parties - which is, in a multipartisan system, the current basilar unit of Brazilian politics.

Using 2014 (country and state-level elections) and 2016 (city-level elections) electoral data, I'd like to simulate how both of those scenarios would play out. How different Congress and the local councils would be from what they are now? I am not talking only about party composition, but also about things like profession, gender, income and race.

Maybe it would look nice in a dot/range plot or even in a slope graph, showing how much each of those aspects would change.

Imagine, in these charts, religions and countries being replaced by things like "white representatives", "black representatives", "women", "men".

dhhxun9umaa0ds9

dg4mt0pxsaahlsn

Details

Possible headline(s):

Data set(s): The Repositório de Dados Eleitorais has detailed data for the elections I wanna meddle with.

Code repository: Here

Possible problems/fears/questions: Worst thing that could happen: since the vote about the reforms is likely happening this week, they might be rejected and then I'd have to say bye-bye to my story.

Work so far

Download the databases for electoral data and merged then with the ones that describe the candidates in detail.

Checklist

This checklist must be completed before you submit your draft.

RodrigoMenegat commented 7 years ago

Update

I simulated the rule changes in the National Congress, but there was no meaningful change, be it on party distribution, representativeness of minorities or anything else. In absolute numbers, as well, the number of congressmen that would not have been elected is slightly under 9% - not much significative.

Nervetheless, I found out that if the electoral system is changed, there will be significantly more impact in state-level legislative houses. In most of cases, the changes is above the one that would have happened in Congress.

I suspect an even bigger change would happen in City Councils, that will also follow the same rule change. I want to simulate that, as well.

Content

states_change

Any changes in direction or topic?

Yes. I am measuring the effects the reform will have in state and city-level legislative houses. It is way bigger than in the National Congress.

Problems/Questions

I want to simulate the simulated change in every city and then group them by size, region and other categories., I will be looking at a huge dataset and don't know if I rave the RAM power to to do it.

Checklist

RodrigoMenegat commented 7 years ago

Final

Hooray, a Lede Piece made it to the cover of the politics section of a major Brazilian newspaper!

Here it is.

page4 (Really low quality screenshot, sorry!)

Project visuals/text

Here is a cheap Google Translation of the story:

If it approves the district rule vote, the National Chamber of Deputies will implement rules that tend to affect state and municipal politics more than Brasília. This is indicated by a simulation made by Folha with data from the 2014 and 2016 elections.

In the current system, a candidate's votes are counted toward him and his party. Vacancies in the Legislature are distributed according to the percentage of votes that each party received.

In the district vote, only the candidates with the largest number of votes are elected.

If these rules were already valid in 2014, approximately 9% of the Chamber of Deputies, or 45 parliamentarians, would not have been elected. In states and municipalities, the effect would be more meaningful.

In 20 of the 27 states there would be bigger changes than in the federal legislature. The most obvious case is that of Acre, where a quarter of the 24 state deputies would not have been elected. In four more states (ES, AP, RR and DF) the change would be over 20%.

According to Emerson Cervi, a doctorate in political science and a professor at UFPR (Federal University of Paraná), one of the reasons is the characteristic of the National Chamber election, which concentrates the best-known candidates with more resources.

"The Chamber of Deputies, in the current system, is similar to a majority rule election. In order to be elected as a deputy, it is necessary to have a lot of money or other political capital, such as being known through the media or the church," says Cervi. "In the State Assemblies, it is possible that candidates rely more the strength of the party," he says.

In Acre, an Assembly with a smaller variety of parties would have been elected: 16 parties have seats today, against 13 in the projected scenario.

It is not only in regions with fewer inhabitants that the scenario would change substantially.

In Rio, the PMDB would have an even greater majority in the Assembly, going from 15 to 24 parliamentarians. The bench of PSOL would be reduced from five deputies elected to only one.

The effect would be similar in Paraná: in 2014, a party not considered traditional, the PSC, obtained the greater representation, with 12 deputies. It would be only the third in the district rule, with six.

However, according to Márcio Carlomagno, a master in political science and a researcher linked to UFPR, projecting the strength of the parties within the district rule voting systen can be misleading, since a change in the rules would also change the way the parties organize themselves internally. Today, well-known names are on the biggest parties, something that may not happen after the change.

"The simulations are made on a past model. The strategies adopted would be completely different. One of the issues that will be changed is the number of candidates. Today the parties have the incentive to launch as many names as possible. If a candidate gets only five votes, there is a gain for the legend. [In the new model], the famous names will be elected, but they may be in small parties, "says Carlomagno.

COUNTIES Municipal council elections would also be subject to the change of rules. Approximately 75% of cities would have more dramatic changes than in the federal legislature.

The city with the biggest change would be Nova Iguaçu (Rio de Janeiro), in the metropolitan region of Rio, where 8 of the 17 elected councilors would not have assumed the position under the district rule.

In eight capitals - Campo Grande (MS), Teresina (PI), Natal (RN), Fortaleza (CE), Manaus (AM), Goiânia (GO) and Belo Horizonte (BH) - the composition of the Municipal Chambers would change by more than 25%.

Other important regional centers, such as Londrina (PR), Uberlândia (MG) and Feira de Santana (BA) would also have alterations of more than a quarter in the city council.

Now, about the visuals: the charts in Folha de S.Paulo are not mine - they were made by their arts desk.

Just for comparison, here's what I had made previously: no map, simple dot plot, the predicted change on the National Congress displayed for reference as a vertical black line.

states_change

Details

Headline: Distritão would affect more the State Assemblies than the National Chamber

Published website version: Here!

Code repository: Here!

Final data set(s): It's too big to upload on GitHub! We are talking about some gb.

What did you find to be the most difficult part of this project?

'Translating' the rules of the electoral system into Python and, again, dealing with five big different databases that were originally split by state.

Are you satisfied with what you produced? Is there anything you would like to change or improve?

Yes! Maybe it's not that appealing visually, but there's a nice story there.

Checklist