Participatory Budgeting and e-Democracy (part 2): the Belo Horizonte case

This post is based on a paper of mine published by the Electronic Democracy Centre (Zurich University) about the experience of the e-Participatory Budgeting of the city of Belo Horizonte. In part 1 of this post I use extracts from a short article by Dan Jellinek (Headstar) and myself that aimed to present a summary of the published paper. At the end, I will add some information contained in the paper about the votes that was not included in the summary article. 

1)     The e-Participatory Budgeting of Belo Horizonte

“Belo Horizonte is the capital of the Brazilian state of Minas Gerais, with a population of just under 2.5 million, of whom 1.7 million are electors. Following the introduction of district PB in 1993, a further Housing Participatory Budgeting (HPB) project was also launched in 1996 to help address an increasing demand for housing in the city. In both the district and housing PB processes, a series of assemblies are held enabling citizens to allocate budgetary resources and scrutinise public spending.

Every two years, the city administration and community leaders invite citizens to the official opening of the PB and to the district rounds in each of the city’s nine districts.

During this first round, the administration distributes a form to neighbourhood representatives to be filled in with citizens’ requests for public works. The representatives in turn call community meetings to establish what the priority public work is for their area. The feasibility of each demand is then technically assessed by the administration.

The administration presents the budget available to each sub-district, which is proportional to a sub-district’s population size and inversely proportional to its quality of life index. The sub-district forums pre-select a maximum of 25 public works for each district, and tours are organised during which the sub-district delegates visit the sites of these works to gain a better understanding of them.

The District Forum is the last deliberative stage of the PB, where the city administration indicates the estimated costs of each of the 25 pre-selected works. Based on these indications and on what the sub-district delegates consider to be priorities, they choose a maximum of 14 works. During this forum the sub-district delegates also elect the district delegates that will follow-up and oversee the execution of the public works. The final stage is the Municipal Meeting of Budgetary Priorities, where the elected delegates present to the mayor the public works selected by the PB to be executed by the administration.

In 2006, alongside the regular PB process explained above, the city administration launched a system of Digital Participatory Budgeting (e-PB). Independent of the budget of US$43 million allocated for the traditional PB, a fund of US$11 million was allocated to the new initiative.

The e-PB consists of a scheme where citizens registered as electors in Belo Horizonte, independent of their place of residency in the city, vote exclusively online for one out of four public works for each of the nine districts of the city. The initiative had three main goals: to modernise the participatory budgeting process through the use of ICTs; to increase citizens’ participation in the process; and to broaden the scope of public works that are submitted to voting (for a Brazilian language site on the project see http://opdigital.pbh.gov.br/ ).

Traditionally, the level of public participation in PB processes had been very low, composed in general of citizens of an advanced age and of lower socio-economic background; in the previous four years only 1.46% of the population participated in the second round of the process. The internet was seen as a way of making it easier for citizens to take part, reducing the time and cost of participation; the traditional PB required citizens to attend meetings at a certain time and place, whereas with the e-PB citizens were free to vote online within a period of 42 days.

For the e-PB, four public works per district were subject to online voting with the aim of selecting one work per district. Citizens over 16 years old were able to vote through an e-voting platform on the city’s website.

In general, the works selected for online voting were much larger than the public works put forward by the traditional PB process. As an example, in the medium-sized district of Barreiro, four choices were offered to voters: to build a new public sports complex; to build a new library; to renew one of the area’s main streets; or to regenerate the district’s commercial centre. Each project was priced at 1.2 million US Dollars and the sports complex won the vote. This is not a process to be taken lightly, since the other three projects did not go ahead.

The e-PB was heavily promoted and the website provided detailed information on the proposed works that were to be selected. Further information could be obtained by email and a designated address was set up to respond to queries. The online platform of the e-PB offered possibilities for multilateral interactivity and, consequently, facilitated deliberative action.

Participation was opened to all citizens, with a discussion forum including nine different threads, one for each district. Even though  active participation in the forum was low, reaching a total of 1,210 posts, all posts could be seen without logging in by all of those who accessed the link to the forums, and the number of readers was significantly higher than the number of posts.

The total number of votes was 503,266 with a total number of 172,938 voters. The difference between the number of voters and number of votes is accounted for by the fact that voters were allowed to vote nine times as long as they voted for only one work per district. These numbers therefore correspond to a participation level of around 10 per cent of electors, nearly seven times more participants than the traditional participatory budgeting (and using a budget nearly seven times smaller).”

2) Analysis of the votes

Now, I would like to add some relevant information that was in the paper and that was not contained in the previous article concerning the votes.

a) Local does seem to matter: As mentioned before, electors were allowed to cast nine votes each (one for each of the 9 districts). Nevertheless, more than half of the voters (52.1%) chose to vote for only one district and nearly two thirds of voters (73.61%) choose to vote for between one and three districts only. Also, qualitative data seems to confirm that citizens preferred to cast for their vote in an informed manner, rather than behaving as free riders that would randomly cast their votes.

b) Absence of socio-economic bias: At least at the aggregate level, no socio-economic bias was found. In this respect, there is no evidence that richer neighbourhoods produced higher levels of participation. Nonetheless, due to the absence of individual level data such analysis should be taken into consideration prudently.

c) Remote voting was essential: The available data shows that at least 30% of the votes were cast from registered electors in Belo Horizonte that, at the moment their vote was cast, were not physically in the city. In other words, it is probable that nearly 1/3 of the voters would not have participated if it hadn’t been for the possibility of casting their votes through the Internet.

I would like to write much more about the discussion contained in the paper and on the feedback that I have received from practitioners and scholars since the publication of the paper, but this post is already too long. What I can say is that, to me, the convergence of PB and ICTs might be one of the most promising and exciting venues for e-Democracy for the years to come: experiences similar to the one of Belo Horizonte are starting to emerge everywhere, and I am convinced that much innovation towards citizen participation will be achieved along this path.

I look forward to the reactions of the colleagues from the Connected Republic, while I contemplate what a participatory budgeting 2.0 should look like.