Back to Stories

Coalition support in Brazil’s Chamber of Deputies

Political appointments
Roll call votes
Legislative Behavior
Published

October 16, 2026

These charts show how coalition parties in Brazil’s Chamber of Deputies voted on executive bills between 2003 and 2018, alongside the positions they held in government.

NoteSource

Aggregated from the replication data for “Bureaucratic appointments and coalition support in presidential systems” (Party Politics, 2026). Only party and coalition totals are shown, and party results based on fewer than five deputies voting are withheld.

The data at a glance

What the replication data contain and how much of it reaches this page.

Support and positions over time

Pick a party and a tier. Shaded columns show the party’s share of posts in that tier for each coalition-year it was a member. Dots show, for each roll call, the share of the party’s voting deputies who followed the president, and the line marks the coalition-year average.

Party
Positions

Share of posts Roll call Coalition-year average Not in the coalition No roll calls in the data

Show as table

Roll calls

Each bar is one roll call on an executive bill, split into the share of deputies voting who followed the president (right) and who did not (left). Choose a party to see its deputies alone, with a tick marking all coalition parties for comparison. Click a dot in the overview to show only that president's roll calls.

Party
Bill type

Followed the president Did not follow (against or abstained)

Show as table

Support across groups

Each dot is one party on one roll call. Most results sit at or near 100%, so the comparison rests on the average and on how often a party fell below 90%.

Compare by
Average Middle half of results, line at the median Party result on one roll call

“Below 90%” is the share of party results in which fewer than 90% of the party's voting deputies followed the president. Party results with fewer than five deputies voting are left out. Dots and averages are not weighted by party size; the pooled column in the table counts every deputy vote equally. Centrão follows the classification in the replication data. Election timing follows the article's indicator for roll calls held after the first round of national or local elections.

Summary table

Parties compared

Parties stay in the same order, from highest to lowest support for the president, so switching the measure shows how shares of posts and ideological distance line up with support.

Show
Centrão party Other party Range of coalition-year averages (3+ roll calls)

Support is the share of the party's coalition votes that followed the president, with the range of its coalition-year averages over periods with at least three roll calls. Shares of posts are means of the party's year totals across the years it voted as a coalition member, with the lowest and highest year. Distance to the president's party is the article's ideological distance measure, averaged across the party's roll calls, with the range across years.

Show as table

Correlates of support

Simple correlations between each party-year’s support for the president and the variables used in the article. Click a bar to see the party-years behind it.

Comparison

Positive correlation Negative correlation

These are simple correlations across party-years (2016 is split by president), without the party fixed effects and controls used in the article's models, so they can differ from the article's estimates. Variables keep the order of their correlation with support across all party-years, whatever the selection. “Within parties” subtracts each party's own average first, so only changes within parties over time remain; centrão does not vary within parties and drops out. Party-years with fewer than five votes are left out.

Show as table

About the data

Every number on this page is computed from the article’s replication data and published only as party and coalition totals. Roll calls come from CEBRAP and appointments from ENAP. The charts are descriptive: they show associations in the data, without the controls used in the article’s models.